Contrib Mineral Petrol (2011) 161:153–173
DOI 10.1007/s00410-010-0518-y
ORIGINAL PAPER
Crystal growth during dike injection of MOR basaltic melts:
evidence from preservation of local Sr disequilibria in plagioclase
Georg F. Zellmer • Kenneth H. Rubin •
Peter Dulski • Yoshiyuki Iizuka •
Steven L. Goldstein • Michael R. Perfit
Received: 3 December 2009 / Accepted: 25 March 2010 / Published online: 10 April 2010
Ó Springer-Verlag 2010
Abstract Profiles of a total of 23 plagioclase crystals
erupted within the 1982–1991 and 1993 flows of the
Coaxial segment of the Juan de Fuca ridge, the 1996 flow
of the North Gorda ridge, and from the Western Volcanic
Zone of the ultra-slow spreading Gakkel Ridge, have been
studied for variations in major and trace element concentrations. We derive equilibration times for the relatively
rapidly diffusing Sr in mid-ocean ridge basalt (MORB)
plagioclase crystals of the order of months to a few years in
each case. All crystals preserve diffusive disequilibria of
Communicated by J. Hoefs.
Electronic supplementary material The online version of this
article (doi:10.1007/s00410-010-0518-y) contains supplementary
material, which is available to authorized users.
G. F. Zellmer (&) Y. Iizuka
Institute of Earth Sciences, Academia Sinica,
128 Academia Road, Section 2, Nankang,
Taipei 11529, Taiwan, ROC
e-mail: [email protected]
G. F. Zellmer S. L. Goldstein
Lamont-Doherty Earth Observatory of Columbia University,
61 Route 9W, Palisades, NY 10964, USA
K. H. Rubin
Department of Geology and Geophysics, SOEST,
University of Hawaii at Manoa, 1680 East West Road,
Honolulu, HI 96822, USA
P. Dulski
Section 3.3, GFZ German Research Centre for Geosciences,
Helmholtz Centre Potsdam, Telegrafenberg,
14473 Potsdam, Germany
M. R. Perfit
Department of Geological Sciences, University of Florida,
Gainsville, FL 32611, USA
strontium and barium. Crystal residence times at MORB
magmatic temperatures are thus significantly shorter, of the
order of days to a few months at most, precluding prolonged crystal storage in axial magma chambers and
instead pointing to rapid crystal growth (up to
*10-8 cm s-1) and cooling (up to *1°C h-1) shortly
prior to eruption of these samples. Growth of these crystals
is therefore inferred to occur almost entirely within oceanic
layer 2 during dike injection. Crystals that grew at lower
crustal levels or earlier in the differentiation sequence
appear to have been excluded from the erupted magmas, as
might occur if most of the gabbroic rocks in oceanic layer 3
formed an interlocking crystal framework, with viscosities
that are too high to carry earlier formed crystals with the
melt. The vertical extent of eruptible, crystal-poor melt
lenses within the gabbroic zone is constrained to *1 m or
less by considering the width of local equilibrium growth
zones, equilibration times, and crystal settling velocities.
This lengthscale is consistent with field evidence from
ophiolites. Finally, crystal aggregates within the Gakkel
ridge sample studied here are the result of synneusis within
the propagating dike during melt ascent.
Keywords Partitioning Diffusion Axial magma chamber Dike injection Synneusis
Introduction
Mid-ocean ridge (MOR) volcanism produces most of
Earth’s surface igneous crust. There is an extensive literature on how aspects of ocean crust creation (melting,
magma accumulation, differentiation, heat loss, and eruption) vary within and between ridge segments. Much
remains to be learned about the details of such processes,
123
154
particularly the range of magmatic conditions associated
with individual MOR eruptions, where spatial geochemical
patterns within individual lava flows may record axial
magma reservoir conditions (e.g. variable melt delivery
and along-axis mixing) as a function of crustal temperature
and magma supply (e.g. Sinton and Detrick 1992; Perfit
and Chadwick 1998; Rubin et al. 2001; Bergmanis et al.
2007; Soule et al. 2007).
Erupted MORB compositions reflect a variable overprint
of differentiation and mixing on mantle-derived primary
liquids (e.g. Elthon 1979; Stolper 1980; Rubin et al. 2009).
Global variations in lava chemistry as a function of geography, spreading rate, and crustal structure constrain how
MOR volcanism samples the mantle, such as melting extent
and melt supply to the crust (e.g. Klein and Langmuir 1987;
Niu and Batiza 1991; Langmuir et al. 1992; Rubin and
Sinton 2007; Niu and O’Hara 2008). All of these processes
modulate chemical signals of mantle heterogeneity and
obscure information about the length scales and magnitudes
of compositional variance in MORB (e.g. Sinton et al. 1983;
Thompson et al. 1985; Langmuir et al. 1986, 1992; Klein and
Langmuir 1987; Niu and Batiza 1991; Sinton et al. 1991;
Grove et al. 1992; Reynolds et al. 1992; Perfit et al. 1994;
Smith et al. 1994; Kelemen et al. 1997a; Michael and Cornell
1998; Herzberg 2004). However, these overprints are also a
useful indicator of the conditions of magma transport and
accumulation at ridges, a series of steps that have collectively been termed the ‘‘Ridge Filter’’ (Rubin et al. 2009).
There are numerous first-order physical features of
spreading centres (e.g. axial morphology, relief, structure,
magmatic, and hydrothermal systems) that relate primarily
to the balance of heat supplied by new magma intrusion
and heat lost by conductive and convective (hydrothermal)
cooling of the seafloor (e.g. Chen and Morgan 1990; Phipps
Morgan and Chen 1993). Because magma supply varies
directly with spreading rate at constant crustal thickness,
i.e. away from hot spots or offsets, many physical attributes
of ridges vary with spreading rate.
MOR magma chambers are commonly considered as
composite magma bodies floored by thick, gabbroic crystal
mushes that are only locally topped by smaller, mostly
liquid melt segregations or melt lenses (Sinton and Detrick
1992). Such melt segregations have been seismically
detected beneath ridges with moderate to high spreading
rates. Two important features of such magma chambers are
that a considerable proportion of magma evolution occurs
at conditions of relatively high crystallinity in the crust,
where the mineralogy of the solid phases is significantly
more mafic than erupted lavas (e.g. Langmuir 1989; Sinton
and Detrick 1992; Ridley et al. 2006; Rubin and Sinton
2007), and that liquid-dominated melt lenses, wherein
more extensive fractionation occurs, exist only fleetingly,
or are long-lived only in regions that are replenished
123
Contrib Mineral Petrol (2011) 161:153–173
frequently enough to counteract the effects of cooling in
the shallow crust. Further, Rubin and Sinton (2007) have
suggested that there may be multiple melt segregations in
the crust, whose number and proportion increase with melt
supply, which results in progressively lower mean magma
temperatures, greater amounts of differentiation, more
variable extents of differentiation, shallower crustal depths,
and greater homogenization of mantle compositional
attributes as melt supply increases.
Relatively little is known about the temporal aspects of
MOR volcanism. Only recently have the timescales of
magma genesis and melt transfer been investigated through
the use of U-series isotopes. Early studies of 230Th-238U
disequilibria in MORB suggested timescales of not more
than *105 years (Condomines et al. 1981). Improvements
in analytical capabilities later allowed the measurement of
shorter-lived isotopes and reduced this timescale to
*103 years based on 226Ra-230Th disequilibria (Rubin and
Macdougall 1988; Sims et al. 2002), and more recently to
as little as a few decades based on 210Pb-226Ra disequilibria
(Rubin et al. 2005).
An understanding of the time scales of crystallization in
young magmatic systems is critical in gaining insights into
the dynamics of the magma ascent, storage, and differentiation, as is evident from studies at volcanic arc and hot
spot settings (e.g. Zellmer and Annen 2008, and references
therein). There are a number of techniques available to
constrain crystallization times in recently erupted volcanic
rocks, including crystal size distribution studies (e.g.
Cashman and Marsh 1988; Marsh 1988; Higgins 1996,
2000; Turner et al. 2003; Morgan et al. 2007), and mineral
isochron dating using 238U-230Th isotopes (e.g. Allègre and
Condomines 1976; Pyle et al. 1988; Condomines 1997;
Heath et al. 1998; Turner et al. 2003; Zellmer et al. 2000,
2008) and 226Ra-230Th isotopes (e.g. Volpe and Hammond
1991; Schaefer et al. 1993; Cooper et al. 2003; Turner et al.
2003; Zellmer et al. 2008; Rubin and Zellmer 2009).
Alternatively, if the temperature of a magma is known, the
residence times of individual crystals may be determined
by modelling the diffusive equilibration of elements within
and between crystals (e.g. Gerlach and Grove 1982;
Nabelek and Langmuir 1986; Humler and Whitechurch
1988; Kohn et al. 1989; Nakamura 1995; Pan and Batiza
2002; Morgan et al. 2004; Costa and Dungan 2005;
Morgan and Blake 2006; Zellmer and Clavero 2006).
Strontium diffusion and magnesium diffusion within plagioclase crystals have recently led to the determination of
residence times of this major mineral phase (Costa et al.
2003, 2010; Zellmer et al. 1999, 2003).
So far there have been two systematic studies of crystal
ages within MORB: Pan and Batiza (2002) used Fe–Mg
interdiffusion to calculate olivine crystal residence times of
hours to years at the East Pacific Rise, with the majority of
Contrib Mineral Petrol (2011) 161:153–173
crystals yielding ages of less than 2 months. They modelled crystal zoning patterns using a continuously fed axial
magma chamber with mean magma residence times of
1–3 months. Conversely, Costa et al. (2010) interpreted
Mg profiles in plagioclase crystals from slow to intermediate spreading ridges and calculated minimum plagioclase
residence times of about 1–10 years. They suggested that
plagioclase crystals resided in a crystal mush zone for years
to decades and were remobilized by disintegration and
entrainment into the magma shortly prior to eruption.
In this contribution, we investigate samples from a
handful of well-characterized MOR eruptions from ridges
of intermediate to ultra-slow spreading rates and apply
strontium diffusion systematics, using previously published
techniques (Zellmer et al. 1999, 2003). In contrast to Costa
et al. (2010), we find ubiquitous short plagioclase residence
times of the order of days to a few months at most, pointing
to rapid crystal growth. These short residence times are
most consistent with crystallization during dike injection
within oceanic layer 2 just prior to eruption.
Geological background and sample description
We have studied plagioclase crystals from two MORB
from the Juan de Fuca Coaxial segment and from one
sample each of the North Gorda Segment and the ultraslow spreading Gakkel Ridge. In this section, the geological background of each of these ridges is reviewed, and the
samples are briefly described. Background geochemical
constraints are provided in Table 1.
Juan de Fuca Ridge, Coaxial segment, and recent
eruptions
The intermediate spreading Coaxial segment of the Juan de
Fuca (JdF) Ridge (*5.8 cm year-1; DeMets et al. 1994)
has a magmatically robust morphology that changes character moving northward away from the Axial seamount
(Embley et al. 1995). In the south, the neovolcanic zone is
dominated by a constructional axial volcanic ridge bisecting a 9-km-wide axial valley, which transitions to a diffuse
series of isolated hills, ridges, and depressions in the north
(Embley et al. 1995). A shallow crustal magma lens was
imaged in the southern part of the segment (Menke et al.
2002). A later seismic experiment imaged an axial magma
chamber (AMC) along several portions of the segment,
getting deeper in its northern part (Carbotte et al. 2006).
The Coaxial segment has had three geochemically distinct (and hence genetically unrelated) eruptions during an
unusually active period between 1981 and 1993 (Chadwick
et al. 1995; Perfit and Chadwick 1998; Smith 1999). Here,
we have studied a sample (2794-2R) from the historical
155
1993 eruption in the northern part of the segment, at a
location called the FLOW site. The eruption lasted several
weeks (Fox et al. 1995) and produced approximately
8.7 9 106 m3 of pillow lava (Perfit and Chadwick 1998).
Two earlier eruptions, one at the FLOW site and one further south at a site know as ‘‘FLOC’’, are presumed to have
occurred sometime between 1981 and 1991 (Embley et al.
2000). We have also studied a sample (2792-4R) from the
earlier FLOW site eruption.
Full geochemical details of the two Coaxial samples
investigated here can be found in Smith (1999). The 1993
flow is more evolved than most older Coaxial lavas and
exhibits moderate internal chemical heterogeneity compared to other mapped MORB lava flows (Rubin et al.
2001). Sample 2794-2R, with *7.9 vol% plagioclase and
*4.6 vol% olivine, is fairly representative of the flow on
average. The earlier FLOW site eruption is slightly less
differentiated. Sample 2792-4R, with *7.8 vol% plagioclase and *5.1 vol% olivine, is again fairly representative
of its flow on average. Samples of both flows also are some
of the most incompatible element depleted of the entire
JdF, and in addition have relatively non-radiogenic Sr
isotopic compositions (Smith 1999; Embley et al. 2000),
indicating that the Coaxial segment has a magmatic system
that is distinct and separate from the other segments of the
JdF ridge (Perfit and Chadwick 1998).
North Gorda Ridge
The intermediate spreading rate of 5.5 cm year-1 of the
North Gorda Ridge is similar to that of the southern Juande-Fuca ridge (DeMets et al. 1994), but North Gorda has a
deep axial valley more typical of ridges at slow spreading
rates (Chadwick et al. 1998). Early petrological and geochemical studies (Davis and Clague 1987; Davis and
Clague 1990) indicated that the North Gorda segment is
characterized by a depleted source and a great diversity of
lava compositions, which was attributed to variations in the
degree of melting and to crystallization and mixing processes within small, isolated magma chambers. More
recently, it has been suggested that the morphology and
increased compositional diversity of North Gorda Ridge
MORB in general reflect a temporal shift in melt production and aggregation under waning melt supply (Davis
et al. 2008). However, lavas from the only documented
historical eruption of this ridge, the 1996 fissure eruption in
the Narrowgate region, are generally more magnesian (i.e.
less differentiated) than most older samples from the region
(Rubin et al. 1998, cf. Table 1). A similar compositional
pattern is observed at the East Pacific Rise at 18.5°S, where
the most recent, higher MgO lavas have been interpreted to
reflect a magmatic rejuvenation on a local portion of the
ridge following a period of lower volcanic productivity and
123
156
Contrib Mineral Petrol (2011) 161:153–173
Table 1 Background geochemical data on studied MORB samples
Sample
Flow
2792-4R
JdF 1982–1991 FLOW site
2794-2R
JdF 1993 FLOW site
W9604-C3
North Gorda 1996
D27-16
Gakkel WVZ
Sample
glassa
Flow av.
(n = 10)a
WRb
Sample
glassa
Flow av.
(n = 24)c
WRb
Sample
glassd
Flow av.
(n = 12)c
WRb
Sample
glasse
Dredge av.
(n = 5)
WRb
SiO2 (wt%)
49.6
49.4
49.4
50.6
50.5
50.3
51.0
51.0
50.7
48.3
49.2
48.2
TiO2 (wt%)
Al2O3 (wt%)
2.02
14.4
2.00
14.5
1.77
14.9
1.65
13.5
1.68
13.6
1.46
14.1
1.23
15.7
1.26
15.8
1.10
16.0
0.88
18.0
1.07
17.1
0.52
24.4
FeO (wt%)
12.3
12.4
11.4
12.7
12.8
11.7
8.20
8.57
7.86
7.85
8.61
4.80
MnO (wt%)
0.20
0.23
0.18
0.23
0.24
0.20
0.14
0.13
0.13
0.12
0.16
0.07
MgO (wt%)
6.95
7.00
8.93
6.67
6.73
8.42
8.26
8.37
9.94
8.84
9.10
5.85
CaO (wt%)
10.8
10.9
10.5
11.3
11.32
11.0
11.8
12.3
11.5
11.8
12.0
13.8
Na2O (wt%)
2.74
2.80
2.64
2.65
2.67
2.59
2.69
2.85
2.56
2.60
2.66
2.33
K2O (wt%)
0.16
0.16
0.14
0.12
0.12
0.11
0.11
0.10
0.10
0.04
0.06
0.03
P2O (wt%)
0.21
0.20
0.18
0.14
0.15
0.12
0.12
0.11
0.11
0.07
0.09
0.04
f
Sr (ppm)
103
ND
ND
91
ND
ND
160
ND
ND
ND
ND
ND
Ba (ppm)
36
ND
ND
25
ND
ND
7.1f
ND
ND
ND
ND
ND
ND not determined
Smith (1999)
a
b
Calculated by mixing glasses with phenocryst phases (olivine Fo90 and plagioclase of average composition at each eruption site) in the
appropriate vol. proportions, which are: ol : plag (vol%) = 5.1 : 7.8 (2792-4R), 4.6 : 7.9 (2794-2R), 4.6 : 6.7 (W9604-C3), 1.15 : 45.1 (D27-16);
and assuming a basalt density of 3 g cm-3 and appropriate phase densities
c
Rubin et al. (2001)
d
Rubin et al. (1998), unless indicated otherwise
e
EPMA average of 1,156 groundmass analyses, this study
f
Unpublished metadata from Rubin et al. (1998)
rifting (Sinton et al. 2002). Recent magmatic activity at the
North Gorda eruption site, which is near the mid-segment
high, may thus reflect a local magmatic rejuvenation that is
superimposed on the waning regional magmatic trend
indentified by Davis et al. (2008).
The 1996 fissure eruption produced approximately
18 9 106 m3 of lava (Chadwick et al. 1998) over a period
of several weeks (Fox and Dziak 1998; Rubin et al. 1998).
The sample investigated here (W9604-C3) is a dated
N-MORB from that eruption and geochemically fairly
representative of the entire flow. It is a fresh, glassy pillow
fragment containing *6.7 vol% plagioclase phenocrysts
and *4.6 vol% olivine microphenocrysts. An earlier
attempt to constrain plagioclase crystallization timescales
in this sample used 226Ra-230Th disequilibrium (Cooper
et al. 2003). That study did not yield an age, but concluded
that crystal growth was rapid over a short time span relative
to the 1.6 kyear half-life of 226Ra.
continental margin of the Laptev Sea in the East (DeMets
et al. 1994). The ridge can be divided into a remarkably
linear western volcanic zone (WVZ), a central sparsely
magmatic zone, and an eastern volcanic zone of widely
spaced volcanoes (Michael et al. 2003). The WVZ segment
has five axial volcanic ridges that are magmatically robust.
The sample studied here (Gakkel D27-16) is part of a very
fresh lava tube with megacrysts (up to *1 cm in size) of
plagioclase feldspar, recovered in dredge 27 by the USCGC
Healy crew on 20 August 2001 during the Arctic Mid-Ocean
Ridge Expedition (AMORE) from the deep between the two
largest volcanic ridges in the centre of the WVZ (cf.
Michael et al. 2003). Due to the porphyritic nature ([45%
crystals) of this sample (D27-16A) and the small amount of
material available, we have used the electron microprobe to
analyse the groundmass composition. The average of 1,156
analyses compares fairly well with an average of five other
MORB samples from this same dredge (cf. Table 1).
Gakkel Ridge, Western Volcanic zone
Methodology review
The ultra-slow spreading Arctic Gakkel Ridge (Dick et al.
2003) lies in the northern extension of the Mid-Atlantic
Ridge (MAR). Its spreading rate ranges from 1.5 cm year-1
near the Lena Trough in the West to 0.6 cm year-1 at the
123
The geospeedometric method applied here is based on the
previous work of Zellmer et al. (1999) and Zellmer et al.
(2003), which is briefly reviewed below. In general,
Contrib Mineral Petrol (2011) 161:153–173
157
calculating crystal residence times at magmatic temperatures involves (1) determination of the chemical equilibrium concentration of a trace element within the
plagioclase crystal under consideration, (2) choosing an
appropriate initial trace element concentration profile, and
(3) applying known trace element diffusion rates within the
crystal to derive a residence time. Here, as previously, we
use Sr as the trace element of choice.
Determination of Sr chemical equilibrium
In chemical equilibrium, the relative Sr concentration of
any two parts i and j of a plagioclase crystal will be
determined by a solid–solid partition coefficient Di/j, which
is a function of the difference in their anorthite contents,
XAn, and of temperature, T:
!
i
j 26; 700 XAn
XAn
i=j
DSr ¼ exp
;
ð1Þ
RT
where R is the universal gas constant (8.31451 J
mol-1 K-1). In practice, absolute bulk crystal equilibrium
profiles are calculated by adopting a boundary condition
such as preservation of total Sr content within the crystal (e.g.
Zellmer et al. 1999, 2003; this study), or chemical
equilibrium of the crystal with an infinite melt reservoir of
the composition of the glass (e.g. Costa et al. 2003, 2010).
However, regardless of which boundary condition is used,
the assessment of Sr chemical equilibrium or deviation
thereof involves comparing Sr concentration ratios of
different parts of a plagioclase crystal (rather than their
absolute Sr concentrations) to the expected chemical
equilibrium ratios derived for these parts. Chemical
equilibrium between two parts i and j of a crystal is
attained when
i
Cobs
j
Cobs
¼
i
Cequil
j
Cequil
¼ Di=j ;
ð2Þ
where C denotes trace element (in this case, Sr)
concentration, implying that in equilibrium,
j
i
Cobs
Cobs
¼
:
i
j
Cequil
Cequil
ð3Þ
If parts i and j are adjacent points within a traverse, local
equilibration may be assessed as follows: We define q as
the ratio between observed and equilibrium concentration
at each point within the traverse, based on any chosen
boundary condition:
q¼
Cobs
:
Cequil
ð4Þ
Local equilibration is attained where the q-profile has no
slope, i.e. when dq/dx = 0, x being distance along the
traverse. It should be noted that distinct crystal growth zone
boundaries are not required to assess local chemical
equilibrium between adjacent parts of a crystal. This is a
major advantage over conventional geospeedometric
approaches, which use diffusion across sharp compositional contrasts to estimate crystal residence times.
In practice, the assessment of local disequilibria requires
knowledge of the uncertainties in both anorthite contents
and trace element concentrations under consideration. It
can be demonstrated (Zellmer et al. 2003) that concentrations Ci and Cj can be considered in equilibrium at the 95%
confidence limit if
i
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
C
Di=j 2 rðCi =C j Þ 2 þðrDi=j Þ2 ;
ð5Þ
C j
where r denotes standard deviation, and
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2ffi
w i=j i
rDi=j ¼
D
rXAn
þ rX j ;
An
RT
ð6Þ
where the constant w is dependent on the element in
question. It was shown by Blundy and Wood (1991) that
wSr = -26,700 ± 1,900 J mol-1 and wBa = -38,200 ±
3,200 J mol-1.
Choosing an appropriate initial Sr concentration profile
The choice of an initial Sr concentration profile is the least
constrained task in determining plagioclase crystal residence times. In conventional geospeedometry, sharp compositional contrasts, e.g. in major element chemistry, are
assumed to have initially also been displayed by the diffusing species, providing a strong boundary condition (e.g.
Costa et al. 2003). In the absence of sharp compositional
contrasts, however, other constraints need to be applied.
For example, extreme initial concentration contrasts across
adjacent crystal growth zones can be employed to derive
residence time maxima (e.g. Zellmer et al. 1999). Alternatively, initial concentration profiles may be modelled if
reasonable assumptions can be made about the petrogenetic
processes operating during crystal growth. For example,
Zellmer et al. (2003) demonstrated that in plagioclase from
the Soufriere Hills andesites, Montserrat, initial Sr concentrations were relatively invariant across phenocrysts
due to the counterbalancing effect of Sr depletion in the
melt during crystallization with Dplagjmelt
[ 1 and increasSr
ing partitioning of Sr into progressively less calcic plagioclase that formed during progressive crystallization
from the evolving melt. Constraints on melt trace element
evolution during crystallization were also employed by
Costa et al. (2010), who used Mg diffusion in plagioclase
to calculate MORB plagioclase residence times and argued
that the evolving melt would be progressively depleted in
Mg due to concurrent crystallization of olivine. However,
123
158
Contrib Mineral Petrol (2011) 161:153–173
as crystallization in their samples was apparently dominated by plagioclase with Dplagjmelt
\1, the melt Mg conMg
tent may in fact have increased with progressive evolution.
As a result, their observed Mg concentration profiles may
have been close to chemical equilibrium initially, rather
than as a result of diffusive equilibration. (A similar case
has previously been made for Ba, cf. Zellmer et al. 2003.)
This example shows that the assumptions made in choosing
an initial concentration profile are critical and need to be
carefully evaluated.
In this contribution, we apply a model of extreme
initial compositional contrasts between adjacent parts of a
crystal to obtain residence time maxima (see the section
on ‘‘Diffusive equilibration of Sr disequilibria’’, below,
for details), and refrain from any assumptions that require
an a priori knowledge of the petrogenetic processes
operating.
preserved (e.g. Zellmer et al. 1999, 2003; Costa et al.
2010), indicating that crystal residence times at magmatic
temperatures are generally much shorter than required for
bulk crystal equilibration. However, one of the strengths of
the methodology applied here is that it allows the assessment of local chemical disequilibria between closely
spaced parts of a plagioclase crystal, irrespective of distinct
growth zone boundaries. Thus, while the entire crystal may
not have attained complete chemical equilibration through
Sr diffusion, some parts of the crystal may already have
equilibrated locally. Yet, if significant chemical disequilibria are retained at a local level in many parts of the
crystal, very short crystal residence times are implied (cf.
Zellmer et al. 2003). Therefore, fine scale analysis potentially allows crystal residence time constraints to be
extended to very short timescales.
Deriving crystal residence times through known Sr
diffusion rates
Analytical techniques
Anorthite profiles are preserved over the time scales considered here, because coupled diffusion of Na, Ca, Al, and
Si is very slow (Morse 1984) compared to diffusion of Sr
(Cherniak and Watson 1994; Giletti and Casserly 1994).
The diffusion coefficient of Sr in plagioclase, ÐSr (where the
use of Ð is to avoid confusion with the partition coefficient,
D), is a function of temperature and anorthite content:
(
)
ÐSr = 10 − (4 .1 X An + 4 .08 ) exp − 3.32 × 10 4 T ,
(7)
where the analytical uncertainty of ÐSr is approximately a
factor of 2 at the ±2r level (Giletti and Casserly 1994).
Zellmer et al. (1999) defined the undiffused fraction d of
initial Sr chemical disequilibrium as
d¼
CSr CSr;equil
;
CSr;0 CSr;equil
ð8Þ
where CSr, CSr,0, and CSr,equil are the measured, initial, and
equilibrium Sr concentrations of a crystal growth zone,
respectively. They also showed that for plagioclase,
d = 10% after a time t & b2/Ð, where b is the half-width
of the growth zone. We will show in this contribution (see
section on ‘‘Equilibration of extreme initial disequilibria in
MORB plagioclase crystals’’, below) that even for unrealistically extreme initial Sr variations, b2/Ð is sufficient time
to achieve equilibration at the 95% confidence limit, as
governed by Eqs. 5 and 6.
Crystals that show Sr concentration profiles in chemical
disequilibrium have cooled to temperatures below which Sr
diffusion is insignificant, before being able to fully equilibrate. In nature, bulk crystal Sr disequilibria are frequently
123
Major and trace element profiles of all samples were
measured by laser ablation inductively coupled plasma
mass spectrometry (LA-ICPMS) at the GeoForschungsZentrum Potsdam, Germany. Details on the instrument,
analytical conditions, experimental procedure, and data
reduction are given in the ‘‘Appendix’’.
Quantitative electron probe microanalysis (EPMA) of
crystals and groundmass and elemental distribution mapping of selected crystals were conducted using a field
emission probe at the Institute of Earth Sciences, Academia
Sinica, in Taipei, Taiwan. All details are provided in the
‘‘Appendix’’.
We have compared EPMA-derived and LA-ICPMSderived anorthite contents (cf. supplementary Fig. S1) and
conclude that the use of LA-ICPMS-derived XAn values is
preferable, because they are based on exactly the same
volume of ablated material, such that artefacts introduced
by compositional heterogeneities within that volume affect
both major and trace elements to a similar degree. We
therefore use the LA-ICPMS-derived XAn values here. For
details, see the ‘‘Appendix’’.
Results
Plagioclase petrography
Major element zonation in the plagioclase crystals studied
here is generally minor, with an average anorthite range for
individual crystals of 0.06 ± 0.04 (1 stdev, n = 23) mole
fraction units. As a result, enhanced elemental maps are
required to illustrate zoning patterns, which are generally
not imaged well enough by backscattered electrons (cf.
Contrib Mineral Petrol (2011) 161:153–173
Fig. 1 Sodium elemental maps
and backscattered electron
images of selected plagioclase
crystals. Within each image,
warmer colours indicate higher
Na concentration, but note that
there is no cross-calibration
between the different images.
Juan-de-Fuca Ridge: a 27942R-1, b 2794-2R-2, c 2794-2R3, d 2792-4R-2, e 2792-4R-5
and 2792-4R-6; Gorda Ridge:
f W9604-C3-4, g W9604-C3-6
amongst other crystals,
h W9604-C3-7 amongst other
crystals; Gakkel Ridge:
i Euhedral plagioclase
megacryst (*1.3 mm across)
from sample D27-16, displaying
complex but concentric zoning.
j A larger ([3 mm across),
subhedral, blocky plagioclase
megacryst from sample D27-16,
pointing to aggregation of
several large crystals into one
compound megacryst. The
lower part of this megacryst
contains microinclusions of
resorbed amphiboles
159
Juan de Fuca Ridge
(b)
(a)
(c)
(e)
(d)
Gorda Ridge
(g)
(f)
(h)
Gakkel Ridge
(i)
Fig. 1). While some crystals are virtually unzoned
(Fig. 1a–c), minor zonation patterns include normal zoning
(Fig. 1d) and complex concentric zoning (Fig. 1e–i). Even
in the case of complex zoning, internal growth surfaces are
generally euhedral. Resorption surfaces are rare and if
observed are generally subhedral and not especially
prominent.
Several of the images show crystals that touch each
other and begin to grow together. The effect of crystals
growing together is most apparent in the large subhedral
Gakkel megacrysts (e.g. Fig. 1j), which represent blocky
(j)
crystal aggregates, composed of several large crystals of
plagioclase with variable contents of microinclusions. The
result is complex non-concentric zoning. These crystal
aggregates are dominantly monomineralic and texturally
very different from polymineralic xenocrysts that commonly display sieve textures caused by internal remelting
(cf. Ridley et al. 2006). Crystal aggregates were avoided
during the microanalytical work undertaken in this study,
which instead targeted simpler, but large, individual crystals of up to 2 mm size, which either displayed concentric
zoning (cf. Fig. 1j) or were not visibly zoned.
123
160
Plagioclase crystal chemistry
The full mineral chemistry of all plagioclase analyses of
the crystals studied here is given in supplementary Table
S1. In summary, the crystals display a relatively restricted
range in anorthite content, with An62-79, An65-72, An70-84,
and An79-84 for the 1982–1991 JdF, 1993 JdF, 1996 Gorda,
and the Gakkel sample, respectively. Plagioclase strontium
and barium concentrations ranges vary, but overlap
between individual crystals of each eruption. Strontium
concentrations of 150–187, 133–159, 139–255, and 161–
264 ppm, and barium concentrations of 0.6–8.0, 2.0–5.6,
0.6–3.4, and 0.4–3.3 ppm were measured in crystals from
the 1982–1991 JdF, 1993 JdF, 1996 Gorda, and the Gakkel
sample, respectively. The lower relative concentration
range in Sr compared to Ba is expected due to the counterbalancing effect of increasing Sr partitioning with
decreasing anorthite content and concomitantly decreasing
melt Sr content, as discussed in detail previously (Zellmer
et al. 2003).
Rim-to-rim geochemical profiles
Given the size of the laser ablation pits, the so called ‘‘rim’’
spot analyses in this study are not representative of the
actual crystal rims, but instead of the composition of the
crystal just inbound of the rim. Therefore, our data do not
provide any information about potential disequilibria
between crystal rims and adjacent glass. Typical XAn, Sr,
and Ba rim-to-rim geochemical profiles of one each of the
crystals from the 1982–1991 JdF, 1993 JdF, 1996 Gorda,
and the Gakkel sample are shown in Figs. 2, 3, 4, and 5,
respectively. The XAn and Sr profiles of the remaining 19
crystals studied here are provided in supplementary Figs.
S2 to S5. Each diagram displays (i) the measured variation
in LA-ICPMS-derived anorthite content constrained by
plagioclase stoichiometry, (ii) the measured variation in Sr
(and Ba) across a crystal, (iii) the expected variation in Sr
(and Ba) at chemical equilibrium as calculated from the
variation in anorthite content (cf. Eq. 1) at the temperatures
given in Table 2, and (iv) the variation of local disequilibria between adjacent parts of the crystal (dq/dx), see the
‘‘Methodology review’’ for details. Zoning patterns are
variable, but large compositional jumps in XAn that would
point to prominent resorption zones are very rare. There do
not appear to be typical zoning patterns for crystals from
individual MORB samples, or consistent differences in
zoning patters between crystals from different eruption
sites.
As the rim spot analyses of an individual crystal are not
always equidistant from the actual crystal rim, they are not
expected to be within analytical error of each other.
Compositional variations within each crystal are large
123
Contrib Mineral Petrol (2011) 161:153–173
enough to account for discrepancies that are occasionally
observed. In any case, it is evident that although local
chemical equilibrium is attained between some adjacent
growth zones, all of the crystals studied do preserve significant local disequilibria, and none of the crystals have
attained complete chemical equilibrium through diffusion
of Sr or Ba. Furthermore, local disequilibria do not systematically decrease towards the crystal cores. Therefore, it
is possible to put constraints on crystal growth rates and
residence times at magmatic temperatures if equilibration
times can be estimated. Finally, the broad rim-to-rim
symmetry of trace element distributions within the two
largest analysed Gakkel crystals suggests that these are
indeed individual phenocrysts rather than crystal aggregates and are therefore suitable for the geospeedometric
work presented in this contribution.
Diffusive equilibration of Sr disequilibria
An upper limit of crystal residence times at magmatic
temperatures can be derived on basis of preservation of
intracrystalline chemical disequilibria of Sr compared to
the time required to reach a totally equilibrated profile. In
this section, we first constrain the applicable diffusion
coefficients for Sr in MORB plagioclase. We then consider
the general requirements for equilibration of unrealistically
extreme initial concentration profiles in MORB plagioclase
crystals. Finally, equilibration time maxima are derived for
each of the crystals studied here.
Applicable Sr diffusion coefficients
Diffusion of trace elements in plagioclase depends on
temperature and anorthite content, with more rapid diffusion at higher temperature and lower XAn. For Sr, this
relationship has been measured and parameterized by
Giletti and Casserly (1994, cf. section ‘‘Deriving crystal
residence times through known Sr diffusion rates’’ above),
and their parameterization is shown graphically in Fig. 6
for a typical MORB magmatic temperature range. Superimposed are arrows indicating the plagioclase compositional evolution by fractional crystallization within dry and
wet (0.5 wt% H2O) MORB during cooling, as calculated
using the MELTS algorithm (Ghiorso and Sack 1995;
Smith and Asimow 2005) on basis of the bulk compositions
of the four samples studied here (cf. Table 1). Note that
liquidus temperatures and plagioclase compositional evolution are largely independent of pressure and oxygen
fugacity, but do change with water content of the melt (for
details, see caption of Fig. 6). It is evident that plagioclase
compositional trends are subparallel to lines of constant
diffusivities. From the maximum and minimum anorthite
Contrib Mineral Petrol (2011) 161:153–173
(a)
JdF-2792-4R- 1
0.74
XAn
161
0.70
0.66
200
0.020
(b)
0.015
160
120
0.005
0.000
80
dρSr/dx (μm-1)
Sr (ppm)
0.010
-0.005
40
-0.010
-0.015
0
(c)
0.030
Ba (ppm)
0.010
0.000
4
-0.010
dρBa/dx (μm-1)
0.020
6
-0.020
2
-0.030
-0.040
0
0
50
100
150
200
250
300
x (μm)
Fig. 2 a Variation in LA-ICPMS-derived anorthite content of
plagioclase crystal JdF 2792-4R-1. Intracrystalline disequilibria of
this crystal are assessed in terms of b strontium and c barium.
Observed Sr and Ba concentrations are given as solid squares. From
the measured anorthite content, a bulk traverse equilibrium profile has
been calculated (solid line with error bars) and can be compared with
the observed concentrations. The ratio q of observed and bulk traverse
equilibrium concentrations provides insights into local disequilibrium
between individual zones: qSr = Srobs/Srequil; qBa = Baobs/Baequil.
Adjacent zones are in local equilibrium when their q values are
identical, i.e. where dq/dx = 0 (solid straight line). From a series of
linear regressions through adjacent q values, the dq/dx profile is
calculated as a measure of local disequilibria (solid stepped line),
together with their 95% confidence limits (dashed envelope). Linear
regressions were chosen such that the errors were minimized. Local
equilibrium is reached where dq/dx = 0 within the uncertainties
provided. Equilibration time and growth rate estimate in Table 2 are
based on the shaded zone of local disequilibrium. Additional Sr
profiles from the 1982–1991 FLOW eruption of the northern Coaxial
segment are given in Fig. S2
contents measured within each of the four samples studied,
the range in Sr diffusivity applicable for the plagioclase
crystals can be deduced, as indicated by the grey
fields in Fig. 6. Although there are small variations
between the samples, a typical diffusion coefficient of
1 9 10-17 m2 s-1 is evidently a good order of magnitude
estimate for all samples.
Equilibration of extreme initial disequilibria in MORB
plagioclase crystals
In the following, we show that for MORB plagioclase,
t & b2/Ð is sufficient time to achieve equilibration at the
2r (*95%) confidence limit, even for unrealistically
extreme initial Sr variations. In MORB, using the
123
162
0.74
XAn
(a)
JdF-2794-2R-6
0.72
0.70
0.68
0.66
0.010
160
(b)
0.008
0.004
0.002
80
0.000
dρSr/dx (μm-1)
0.006
120
Sr (ppm)
Fig. 3 a Variation in LAICPMS-derived anorthite
content of plagioclase crystal
JdF 2794-2R-6. Intracrystalline
disequilibria of this crystal are
assessed in terms of b strontium
and c barium. Symbols and
notation as in Fig. 2. Additional
Sr profiles from the 1993
FLOW eruption of the northern
Coaxial segment are given in
Fig. S3
Contrib Mineral Petrol (2011) 161:153–173
-0.002
40
-0.004
-0.006
0
(c)
0.025
4
0.020
0.010
0.005
2
0.000
dρBa/dx (μm-1)
Ba (ppm)
0.015
3
-0.005
1
-0.010
-0.015
0
0
50
100
150
200
250
x (μm)
parameterization of Blundy and Wood (1991) and appropriate formation temperatures, DSr ranges from *2.7 for
XAn = 0.6 to *1.4 for XAn = 0.85. An unrealistically
extreme initial disequilibrium profile may thus be envisaged for a hypothetical crystal with alternating overgrowth
zones generated by crystallization from cool E-MORB and
hot N-MORB, with initial Sr concentrations alternating
between 419 ppm within XAn = 0.6 (domain i) and
126 ppm within XAn = 0.85 (domain j), using the above
partition coefficients and typical E-MORB and N-MORB
Sr concentrations (Sun and McDonough 1989). In this case,
taking conservative analytical uncertainties of 1.5% for
XAn and 1.2% for CSr, we obtain for substitution into Eq. 5:
rðCi =C j Þ ¼ 0:056; rDi=j ¼ 0:069 (cf. Eq. 6); and thus
Sr
Sr Sr C i
Srj Di=j 0:18 (cf. Eq. 5) for domains i and j to be in
Sr C
Sr
123
equilibrium at the 95% confidence limit at the low temi=j
perature end of crystal growth (*1,050°C, i.e. DSr ¼ 1:83
as calculated using the above anorthite contents of the two
domains and the parameterization of Blundy and Wood
1991, cf. Eq. 1).
At this low temperature end of crystal growth, equilibrium Sr concentrations are 352 within domain i and
193 ppm within domain j of this hypothetical crystal. Thus,
after a time t & b2/Ð, when initial disequilibria have
decreased by 90% (cf. section ‘‘Deriving crystal residence
times through known Sr diffusion rates’’), Sr concentrations of 359 and 186 ppm would be observed within low
and high anorthite domains of the crystal. We obtain
C i
i
CSr
i=j Sr
¼ 1:93, and hence C j DSr ¼ 0:10\0:18, finding
CSrj
Sr
that the condition for equilibrium is in fact met well before
Contrib Mineral Petrol (2011) 161:153–173
0.85
(a)
Gorda W9604-C3-4
0.81
XAn
Fig. 4 a Variation in LAICPMS-derived anorthite
content of plagioclase crystal
Gorda W9604-C3-4.
Intracrystalline disequilibria of
this crystal are assessed in terms
of b strontium and c barium.
Symbols and notation as in
Fig. 2. Additional Sr profiles
from the 1996 North Gorda
eruption are given in Fig. S4
163
0.77
0.73
0.69
0.012
300
(b)
0.010
0.006
Sr (ppm)
200
0.004
0.002
dρSr/dx (μm-1)
0.008
0.000
100
-0.002
-0.004
-0.006
0
(c)
0.040
3
Ba (ppm)
0.020
2
0.010
0.000
1
dρBa/dx (μm-1)
0.030
-0.010
-0.020
0
0
50
100
150
200
250
-0.030
300
x (μm)
time t & b2/Ð. To be conservative, we refer to b2/Ð as the
equilibration time.
Equilibration times and residence times of the crystals
investigated
Typical timescales to reach a totally equilibrated profile
can be calculated for the crystals studied by appropriate
choice of growth zone width (2b). Growth zone spacing is
not immediately apparent from the profiles. Here, we use
the minimum spacing over which local disequilibria are
preserved within the central part of the rim-to-rim profiles
obtained in this study, marked in grey in Figs. 2, 3, 4, and
5, and S2 to S5. Local disequilibria are calculated following Zellmer et al. (2003), as reviewed in the section on
‘‘Determination of Sr chemical equilibrium’’, above: the
ratio between observed and equilibrium concentration is
defined as q, and local disequilibria are identified in zones
of the crystal where dq/dx = 0. Note that in order to derive
growth zone spacings from Figs. 2, 3, 4, and 5, and S2 to
S5, we plot dq/dx itself, rather than its absolute value |dq/dx|,
which was employed by Zellmer et al. (2003) as a means
of identifying the magnitude of local disequilibria.
Crystal residence times to achieve complete equilibration are of the order of months to a few years (*0.6–
14 years for JdF2792-4R, *0.5–6.1 years for JdF2794-2R,
*0.5–2.5 years for Gorda W9604-C3, and *2.2–34 years
for Gakkel D27-16, cf. Table 2). These equilibration times
are extreme overestimates because they are based on an
unrealistically extreme initial disequilibrium profile as
outlined in the previous section. Hence, more realistic
timescales for crystals to fully equilibrate are likely shorter.
123
164
0.88
(a)
Gakkel D27-16-E
XAn
0.84
0.80
0.004
300
(b)
0.003
0.002
Sr (ppm)
200
0.001
dρSr/dx (μm-1)
Fig. 5 a Variation in LAICPMS-derived anorthite
content of plagioclase crystal
Gakkel D27-16-E.
Intracrystalline disequilibria of
this crystal are assessed in terms
of b strontium and c barium.
Symbols and notation as in
Fig. 2. Additional Sr profiles
from this sample of the Gakkel
Ridge are given in Fig. S5
Contrib Mineral Petrol (2011) 161:153–173
0.000
100
-0.001
-0.002
0
(c)
0.050
2
Ba (ppm)
0.030
0.020
1
0.010
dρBa/dx (μm-1)
0.040
0.000
-0.010
0
0
600
1200
-0.020
1800
x (μm)
However, the fact that local chemical disequilibria are
preserved in all of the crystals studied here suggests that
more realistic, shorter equilibration times still significantly
overestimate actual crystal residence times. We would
therefore argue that actual crystal residence times are likely
at least one order of magnitude shorter than the calculated
equilibration times, i.e. generally ranging from days to a
few months at most. Longer equilibration and residence
times are only possible if crystals resided at much lower
temperatures than those calculated from the least calcic
growth zones. Strikingly, the observed ranges in crystal
residence times are very similar in all the ridges studied
and do not appear to be strongly dependent on differences
in spreading rates or melt supply.
123
Discussion
Crystallization during dike injection
The preservation of significant local disequilibria as evident from Figs. 2, 3, 4, and 5, and S2 to S5 indicates that
the crystals studied here had an order of magnitude shorter
residence times at magmatic temperatures than the calculated equilibration times. Further, the fact that local disequilibria do not systematically decrease towards the
crystal cores suggests that crystal growth was too rapid to
be resolved by Sr diffusive equilibration, even locally. As
shown above, it is hence evident that crystals grew rapidly
and on the order of days to at most a few months prior to
Contrib Mineral Petrol (2011) 161:153–173
165
Table 2 Determining minimum growth rates for all studied crystals
Crystal
XAn range
Min.
T (°C)a
ÐSr (cf. Fig. 6)
(10-17 m2 s-1)
JdF 2792-4R-1
0.64–0.72
1,080
*1
JdF 2792-4R-2
0.67–0.79
1,080
JdF 2792-4R-4
0.62–0.72
1,080
JdF 2792-4R-5
0.67–0.79
JdF 2792-4R-6
JdF 2792-4R-7
Zone width
2b (mm)b
Equil. time
b2/ÐSr (years)c
Distance to
rim (mm)
31
0.8
149
6.2
*1
94
7.0
88
0.4
*1
132
14
96
0.2
1,080
*1
27
0.6
94
5.2
0.68–0.77
1,080
*1
29
0.7
180
8.6
0.69–0.74
1,080
*1
28
0.6
94
4.8
JdF 2794-2R-1
JdF 2794-2R-2
0.65–0.71
0.66–0.68
1,100
1,100
*1
*1
35
30
1.0
0.7
154
156
5.0
6.9
JdF 2794-2R-3
0.67–0.72
1,100
*1
33
0.9
139
5.1
JdF 2794-2R-4
0.67–0.69
1,100
*1
88
6.1
107
0.6
JdF 2794-2R-5
0.67–0.72
1,100
*1
26
0.5
79
4.7
JdF 2794-2R-6
0.67–0.70
1,100
*1
25
0.5
135
8.6
JdF 2794-2R-7
0.67–0.71
1,100
*1
30
0.7
268
11.9
Gorda W9604-C3-1
0.79–0.82
1,130
*1
46
1.7
80
1.5
Gorda W9604-C3-2
0.75–0.80
1,130
*1
56
2.5
132
1.7
Gorda W9604-C3-3
0.73–0.76
1,130
*1
25
0.5
100
6.4
Gorda W9604-C3-4
0.70–0.83
1,130
*1
25
0.5
147
9.4
Gorda W9604-C3-5
0.70–0.72
1,130
*1
26
0.5
69
4.1
Gorda W9604-C3-6
0.71–0.81
1,130
*1
34
0.9
115
4.0
Gorda W9604-C3-7
0.76–0.84
1,130
*1
33
0.9
189
6.9
Gakkel D27-16-E
0.81–0.84
1,145
*1
53
2.2
805
11.5
Gakkel D27-16-H
Gakkel D27-16-O
0.79–0.84
0.81–0.84
1,145
1,145
*1
*1
60
206
2.9
34
973
460
10.8
0.4
a
Min. growth rate
(10-10 cm s-1)d
Based on the lowest anorthite content of all crystals in each rock sample, and assuming 0.5 wt% H2O in the primitive magma
b
This is the width of the disequilibrium zone marked in grey in Figs. 2, 3, 4, and 5, and S2 to S5
c
Actual crystal residence times are likely at least one order of magnitude lower, see text for discussion
d
These are effective growth rates that do not account for limited intermittent dissolution. Actual growth rates are likely at least one order of
magnitude greater, see text for discussion
eruption on the ocean floor. These short crystal residence
times compare to estimates of days to a few years for
olivines in MORB from the Mid-Atlantic Ridge and the
East-Pacific Rise, based on Ni diffusion and Fe–Mg
interdiffusion (Nabelek and Langmuir 1986; Pan and
Batiza 2002) and are compatible with short magma differentiation and transport times required for the preservation of 210Pb-226Ra disequilibria that are generated during
mantle melting (Rubin et al. 2005). They are, however,
significantly shorter than recently reported minimum
MORB plagioclase residence times of about 1–10 years
(Costa et al. 2010).
Based on the calculated equilibration times and the
distance of central zones of local disequilibria from their
crystal rims, we obtain minimum plagioclase crystal
growth rates of the order of 10-9–10-11 cm s-1 (cf.
Table 2). An order of magnitude shorter crystal residence
times indicate growth rates of up to 10-8 cm s-1, implying
cooling rates of up to 1°C h-1 (cf. Cashman 1993).
Cooling rates upon eruption within submarine lava flows,
where rapid quenching prevents significant crystal overgrowth and produces glassy or microcrystalline groundmasses such as those in which the crystals studied here are
found, are typically at least two orders of magnitude faster
(e.g. Grove 1990). Conversely, cooling rates within the
gabbroic lower crust (oceanic layer 3) are four to six orders
of magnitude slower (VanTongeren et al. 2008). Cooling
rates of up to *1°C h-1, yielded by the crystals studied
here, are most consistent with crystal growth occurring
during dike injection (cf. Cashman 1993). One caveat to
this scenario might be the inclusion of older crystals stored
for long periods of time at significantly lower temperatures.
Each 100°C decrease in temperature would result in a
decrease in strontium diffusivity of about one order of
magnitude. Thus, if the crystals studied here were more
than a few years old, they would have had to be stored
at C 200 degrees lower temperatures for most of that time.
Even in the hydrous case, crystallinity would in this case
123
166
Contrib Mineral Petrol (2011) 161:153–173
XAn
0
0.2
0.4
XAn
0.6
0.8
0
0.2
0.4
0.6
0.8
1
1300
1300
(a) JdF2792-4R
(b) JdF2794-2R
1200
T (°C)
H2O
wt%
0.5
m2
s -1
1100
m2
s -1
10 -1
8
m2
s -1
7
10 -1
10 -1
6
m2
s -1
m2
s -1
10 -1
5
0.5
10 -1
8
m2
s -1
7
10 -1
10 -1
6
m2
s -1
10 -15
1100
wt%
m2
s -1
H2O
T (°C)
dr
y
dr
y
1200
1000
1000
(d) Gakkel D27-16
0.5 w
t% H
2
O
dry
(c) Gorda W9604-C3
1200
T (°C)
1100
m2
s -1
8
m2
s -1
10 -1
7
10 -1
10 -1
m2
s -1
10 -1
6
m2
s -1
8
m2
s -1
7
10 -1
10 -1
6
m2
s -1
10 -1
10 -1
5
5
1100
m2
s -1
m2
s -1
0.5
wt%
T (°C)
H2O
dr
y
1200
1000
1000
0
0.2
0.4
0.6
0.8
XAn
0
0.2
0.4
0.6
0.8
1
XAn
Fig. 6 Evolution of anorthite content calculated by MELTS fractional crystallization modelling of MORB whole rock compositions
(Ghiorso and Sack 1995; Smith and Asimow 2005), superimposed on
diffusion coefficients as parameterized by Giletti and Casserly (1994).
MELTS modelling used the FMQ-1 oxygen buffer, as appropriate for
MORB evolution (cf. Christie et al. 1986; Bezos and Humler 2005),
and employed dry MORB versus 0.5 wt% H2O in the parental melt to
cover the range of H2O contents observed within MORB with up to
0.2 wt% K2O (cf. Almeev et al. 2008). Solid circles indicate the
highest and lowest anorthite content observed within each sample,
and symbol sizes represent ±2r uncertainties. Small discrepancies
between the observed and predicted highest anorthite content,
particularly in (a) and (c), may be related to flow heterogeneities
(Rubin et al. 2001), with the samples studied not representative of the
least evolved parts of each flow
have exceeded 70% at the magmatic compositions of the
studied basalts (Kelemen and Aharonov 1998), thereby
effectively forming an uneruptable viscous crystal mush
(e.g. Lejeune and Richet 1995) containing plagioclase
crystals of much lower anorthite content (AnB50) than those
observed in the samples studied here. Although more sodic
crystal rims may have been resorbed prior to eruption due
to immersion of such crystals into hotter, more mafic
magma, it is difficult to envisage a scenario that would
remobilize crystals from such a mush zone without at least
preserving some crystal clots with more sodic plagioclase,
or evidence for internal melting such as sieve textured
crystals or prominent resorption zones. We therefore consider this scenario as highly unlikely.
Our findings appear to be in contrast with the significantly longer minimum plagioclase residence times of
years to decades in samples from the Mid-Atlantic Ridge
and the Costa Rica Rift proposed on basis of Mg profiles,
which were interpreted to reflect crystal residence in and
remobilization from a lower crustal mush zone (Costa et al.
2010). However, as indicated above, the assumption that
Mg disequilibria were introduced during crystal growth, on
123
Contrib Mineral Petrol (2011) 161:153–173
which those residence times are based, needs to be carefully considered. If crystals grew closer to Mg chemical
equilibrium than assumed, the residence times calculated
for these crystals on basis of intra-crystalline diffusion may
be overestimates. A detailed re-evaluation of those published data is beyond the scope of this study, and remobilization of selected crystals from a mush zone remains
possible at some sites. However, significant local Sr disequilibria are also observed within the plagioclase crystals
studied by Costa et al. (2010), which would indicate that
the residence times of at least some crystals erupted from
those ridges may be shorter than previously thought.
Effects on erupted phenocryst size
The size of phenocrysts, once nucleated, will be controlled
by the time available for crystal growth, which depends on
magma rise rate and ascent distance. Although magma flow
is controlled by dike width and melt viscosity (Rubin 1995),
the widths of dikes at oceanic spreading centres typically
range from 0.5 to 1.5 m and do not vary consistently with
spreading rate (Qin and Buck 2008). Similarly, no systematic variations in MOR melt viscosities are expected.
Thus, variations in the length of the melt ascent path are
anticipated to perceptibly affect the size of large, early
nucleated phenocrysts. Globally, Rubin and Sinton (2007)
have shown that with decreasing overall melt supply to midocean ridges (based on spreading rates), the depth of axial
magma lenses increases, implying that the dike section
concomitantly thickens. This is consistent with geophysical
observations of variations in axial magma chamber depth
along individual ridge segments of the Juan de Fuca Ridge
(Canales et al. 2005). Thus, if phenocrysts indeed grow
mainly during diking, then the size of the large crystals
studied here would be expected to broadly anti-correlate
with melt supply, due to the greater magma ascent path. It is
noteworthy that samples from the fast spreading East
Pacific Rise were not studied here because of their general
lack of sufficiently large plagioclase crystals and their
commonly aphyric nature. In the areas where our basalts
were recovered, the North Gorda eruption took place near
the mid-segment high of this intermediate spreading ridge,
implying relatively high melt supply, and crystals are generally small. The northern end of the JdF Coaxial segment,
where the two FLOW eruptions were extruded, has an
isolated hill-ridge-depression morphology, implying relatively lower melt supply, and crystals reach a slightly larger
size. The largest crystals are present in the sample taken
from the deep between two volcanic ridges of the ultra-slow
spreading Gakkel ridge (cf. Table 2). These observations
are thus consistent with crystal growth occurring dominantly during dike injection, rather than within axial magma
chambers of the gabbroic oceanic layer 3.
167
Constraints on conditions within the upper gabbroic
layer
In the previous sections, we have argued that crystal
growth occurred during dike injection, implying that melts
extracted from the gabbroic layer are essentially aphyric.
However, we cannot preclude the existence of microlites
carried by these melts at the onset of diking. The width of
zones within local Sr equilibrium occasionally observed in
the crystals studied here is typically around 50 lm and
rarely exceeds 100 lm. While most of these zones likely
grew in equilibrium, rather than having equilibrated by
diffusion, a small proportion might represent microlites
that were suspended within magma lenses of the gabbroic
zone for sufficiently long to have attained equilibrium, and
that were then used as nuclei for more rapid overgrowth
prior to eruption. We can use the 100-lm-length scale to
put constraints on the vertical extent of such magma lenses
through Stokes’ Law of particle settling:
cs ¼
2 gDq 2
a ;
9 qm m
ð9Þ
where cs is the terminal settling velocity of a spherical
crystal, g is the acceleration due to gravity, Dq is the
density contrast between the crystal and the magma, a is
the diameter of the crystal, and qm and m are the density and
kinematic viscosity of the melt, respectively. For anorthitic
plagioclase in tholeiitic melts, Dq & 50 kg m-3,
qm & 2,700 kg m-3 (Huppert and Sparks 1980), and m
may be as low as 10-2 m2 s-1 (Martin and Nokes 1989),
yielding plagioclase settling velocities in basalts of
B4 9 10-8 m s-1 for small crystals of B100 lm diameter
(cf. Martin and Nokes 1989). Even in anorthitic plagioclase
crystals, chemical equilibrium between two adjacent
50 lm growth zones (to total a 100 lm nucleus) will have
been well attained within less than 3 years (cf. Gakkel
samples E and H in Table 2). After this time, the small
crystals considered above will have settled less than 4 m.
Larger crystals settle faster and would have settled out of
the magma lens into the viscous mush zone. The absence of
equilibrium crystal growth zones of [100 lm width in the
crystals studied here indicates that once removed into the
mush, individual crystals are not commonly remobilized
into the melt and may only be erupted as part of cumulate
xenoliths. However, the likely vertical dimension of melt
lenses within the gabbroic oceanic layer 3 will be significantly smaller than the estimated 4 m because (i) maximum equilibration times in Table 2 overestimate actual
equilibration times as outlined above, and (ii) convection
within the inner part of a melt lens, if present, would
decrease the rate with which crystals may be removed from
the melt through a stagnant boundary layer (Martin and
Nokes 1989). Hence, the vertical extent of the melt lenses
123
168
(c)
Contrib Mineral Petrol (2011) 161:153–173
Our data do not rule out that magmas from such small melt
lenses were subsequently aggregated into a larger body just
prior to eruption. However, if this was the case, aggregation must have been very rapid and residence in the axial
chamber very short.
Origin of complex zoning and crystal aggregates
(b)
Given the limited lateral extent of individual sills within
the layered gabbros (Ernewein et al. 1988; Boudier et al.
1996) and MORB lava flow volumes of several million
cubic metres, melts from more than one sill must be
combined in order to provide enough magma to sustain
eruptions (Fig. 7a). The flow regime within the propagating
dike will depend on the Reynolds number, Re. Spence et al.
(1987) have shown that during buoyancy-driven magma
ascent, the flow will be laminar if
Re ¼
(a)
Fig. 7 Petrogenetic model of melt extraction and magma ascent.
a Initially aphyric melt from several melt lenses within the layered
gabbros combine as diking is initiated. b Crystals begin to grow
during laminar magma flow within the dike. Transfer of crystals
across boundaries of melts with slightly different petrogenetic
histories leads to complex zoning. c Grain dispersive pressures lead
to crystal aggregation by synneusis in the centre of the dike if crystal
content is high enough
from which magmas were drained during the eruptions of
the samples studied here probably did not exceed *1 m.
This latter result is consistent with the dimension of sills
within the layered cumulates of the Oman ophiolite, which
have sill thicknesses ranging from centimetres to decimetres. Metre-sized layers are very scarce (cf. Ernewein et al.
1988; Boudier et al. 1996). The data from our study sites
hence support models of lower oceanic crust formation by
multiple sill intrusion (Kelemen et al. 1997b; Kelemen and
Aharonov 1998; Kawamura et al. 2005) and are consistent
with the notion that there exists a number of poorly connected melt lenses in the gabbroic crust beneath MORs
(Rubin and Sinton 2007), rather than a single axial magma
chamber as originally envisaged (Sinton and Detrick 1992).
123
ch
\103 ;
m
ð10Þ
where c is the dike propagation speed and h is the dike
width. For reasonable values (c = 1 m s-1, h = 1 m,
m C 10-2 m2 s-1), the flow regime will thus be laminar.
Laminar flow is not as conductive to magma mixing as
turbulent flow, and this may be the reason for some of the
observed chemical heterogeneities of sampled MORB lava
flows. Inefficient mixing may also result in the complex
zoning observed in some of the phenocrysts studied here, if
crystals are transferred between melt batches with different
origins and storage histories during magma ascent
(Fig. 7b).
As crystallization progresses, grain dispersive pressures
will lead to preferential migration of phenocrysts towards
the centre of the dike, where velocity gradients are lowest
(Komar 1972). When phenocryst concentration reaches a
value of *8%, crystals in the centre of the dike begin to
interact mechanically, and this may lead to phenocryst
aggregation within the ascending melt by capillarity
(Fig. 7c). This process is known as synneusis (Vance 1969)
and is characterized by like minerals combining preferentially (Vance and Gilreath 1967). While some degree of
phenocryst aggregation is observed in all of the samples
studied here (cf. Fig. 1), synneusis appears to be most
pronounced in the porphyritic Gakkel ridge sample with its
largely monomineralic plagioclase aggregates. These are
texturally very different from polymineralic xenoliths,
which commonly display sieve textures caused by internal
remelting (cf. Ridley et al. 2006). In addition, the change in
crystal habit from prismatic to blocky with increasing
crystal size of Gakkel ridge plagioclase is consistent with
this process. Although Dowty (1980) has questioned if
synneusis represents the principal mechanism in the formation of crystal aggregates in general, there is more
Contrib Mineral Petrol (2011) 161:153–173
recent evidence for it to occur in experimentally cooled
basaltic melts (Duchene et al. 2008; Pupier et al. 2008).
Crystal aggregation is not as pronounced in the less porphyritic samples studied here, such that significantly lower
phenocryst contents (\13% vs. [45% at Gakkel) likely
resulted in less frequent mechanical interactions of
phenocrysts during dike injection.
169
Science Foundation provided generous support for the Alvin and
Jason dives. The research was funded by grants of the National
Science Council of Taiwan (NSC 95-2116-M-001-006, 96-2116-M001-006, and 97-2628-M-001-027-MY2) and the Institute of Earth
Sciences, Academia Sinica, to GFZ. Furthermore, partial support was
provided by NSF grants to KHR (OCE-0732761 and OCE-9905463)
and MRP (OCE-0221541 and OCE-9530299).
Appendix
Conclusions
Laser ablation inductively coupled mass spectrometry
(i)
Plagioclase crystals erupted in MORB from intermediate to ultra-slow spreading ridges preserve significant intracrystalline chemical disequilibria for
strontium and barium, implying crystal residence
times of the order of days to at most a few months at
magmatic temperatures, irrespective of spreading
rate.
(ii) Rapid crystal growth rates of up to 10-8 cm s-1
imply rapid cooling rates of up to 1°C h-1 and
suggest that crystallization occurs during magma
ascent in dikes, rather than within thermally buffered
axial magma chambers.
(iii) At lower melt supply, melt lenses are situated
deeper, resulting in longer magma ascent times and
the growth of larger crystals within erupted basalts
(e.g. Gakkel). Higher melt supply, shallower melt
lenses, and shorter magma ascent times limit crystal
growth (e.g. EPR).
(iv) Crystals larger than 100 lm may grow within the
gabbroic layer but are not usually erupted, implying
that they are retained in a rigid crystal mush. Rapid
crystal settling velocities in basalts suggest that the
vertical extent of axial melt lenses is unlikely to
exceed *1 m in order to preclude incorporation of
microphenocrysts into ascending magmas drained
from such melt lenses. This result is consistent with
field evidence of thin gabbroic sills exposed in the
lower crustal section of the Oman ophiolite.
(v) Complex zoning pattern in some crystals are due to
phenocryst transfer during dike injection between
mingling melts with different storage histories.
Crystal aggregates in the porphyritic Gakkel ridge
sample likely result from synneusis within the centre
of the propagating dike during magma ascent.
Acknowledgments GFZ acknowledges insightful communications
with Daniel Morgan, Matthew Smith, and Steve Sparks, and is
grateful to Jörg Erzinger for his hospitality and support at the
GeoForschungZentrum. Earlier versions of this paper have been
significantly improved by the constructive comments of Fidel Costa
and two anonymous reviewers. We thank Michael Wiedenbeck for
access to the AMNH feldspar standard during the analytical work in
Potsdam. The NOAA Undersea Research Program and the National
Major and trace element plagioclase rim-to-rim profiles of
all samples were measured at the GeoForschungsZentrum
Potsdam using a GEOLAS M Pro (Coherent, Germany)
laser ablation device coupled to an ELAN DRC-e (PerkinElmer SCIEX, Canada) ICP-MS in dual detector mode.
The LA device consisted of an excimer laser (COMPexPRO 102, Argon Fluoride 193 nm, Lambda Physik, Germany) with a maximum output energy of 200 mJ per pulse
for repetition rates between 1 and 10 Hz, an optical beam
path homogenizer to provide a homogeneous laser beam,
an aperture mask with 10 circular holes with diameters of
0.125–4 mm (for ablation pit diameters of 5–160 lm), a
petrographic microscope (Olympus BX 51), a 20 cm3
sample cell and a computer controlled x,y,z-stage. Samples
were observed through the petrographic microscope with
two lenses of different magnification (5 and 20 fold).
During ablation, the sample image was continuously
observed on a separate high-resolution monitor through a
Schwarzschild objective (25-fold magnification) and a
high-resolution CCD camera. Helium was used as carrier
gas (1.0 L min-1), and argon as plasma (15 L min-1),
auxiliary (1.2 L min-1), and nebulizer gas (0.7 L min-1).
Elements of interest were measured using an output
energy of 100 mJ, with 10 ms dwell time and 3 ms quadrupole settling time. Background intensities from the gas blank
were measured for 30 s (laser firing, shutter closed) followed
by acquiring transient signals of the analytes for 60 s (laser
firing, shutter open) or until the laser beam had drilled
through the sample and reached the sample holder (in case of
thin sections). Each analytical batch consisted of up to 20
analyses. Samples were measured using a repetition rate of
5 Hz, an energy density of 6 J cm-2 and spot sizes of 16 or
24 lm. Rim-to-rim geochemical profiles were obtained for
each crystal studied by spot analyses spaced between *20
and *100 lm apart, depending on crystal size. A reference
glass standard (NIST 610) was measured twice at the
beginning and twice at the end of an analytical batch for
external calibration, using a repetition rate of 10 Hz, an
energy density of 10 J cm-2, and a spot size of 32 lm.
For data reduction, the LOTUS 123 macro-based
spreadsheet programme LAMTRACE was employed. A
123
170
short, general description of its capabilities is given by van
Achterbergh et al. (2001), and its algorithms are described
in Longerich et al. (1996). The programme performs
background correction, correction for instrumental drift,
internal calibration, choice of integration intervals, and
calculation of element concentrations using external calibration. The Ca signal was used for the calibration of trace
element concentrations. Repeat analyses of gem labradorite
AMNH 95557 revealed that LA-ICPMS-derived Na/Ca
ratios were consistently about 11% higher than those
derived by electron probe microanalysis (EPMA). This
discrepancy was corrected for in the calculation of LAICPMS-derived anorthite contents, which was based on
plagioclase stoichiometry (cf. Zellmer et al. 2003). The
same repeat analyses also provide the basis for assessment
of relative uncertainties for individual elements, and these
are 1.5% for XAn, 1.2% for Sr, and 2.6% for Ba (1r,
n = 16, cf. supplementary Table S1 for other elements).
Electron probe microanalysis
Following the LA-ICPMS analysis, the specimens were repolished using a vibration polisher with 0.3 lm particle
alumina compounds to remove the debris of LA, and then
carbon coated for microprobe analysis. Quantitative
chemical analyses and, in some cases, elemental distribution (mapping) analyses were made by a field emission
electron probe micro analyser (FE-EPMA: JEOL
JXA-8500F) equipped with five wave-length dispersive
spectrometers (WDS) at the Institute of Earth Sciences,
Academia Sinica, in Taipei. Secondary- and back-scattered
electron images were used to guide the analysis on target
positions of minerals. A 2-lm defocused beam was operated for quantitative mineral analysis at an acceleration
voltage of 12 kV with a beam current of 5 nA. Groundmass
analyses were undertaken on the same instrument with a
5 lm beam diameter, 15 kV accelerating voltage, and 10
nA beam current.
The measured X-ray intensities were ZAF-corrected
using the standard calibration of synthetic chemical-known
standard minerals with various diffracting crystals, as follows: wollastonite for Si with TAP crystal and Ca with
PET crystal, rutile for Ti (PET), corundum for Al (TAP),
chromium oxide for Cr (PET), fayalite for Fe with LiF
crystal, tephroite for Mn (PET), pyrope for Mg (TAP),
albite for Na (TAP), and adularia for K (PET). Counting
times for peak and both upper and lower baselines were
20 s and 10 s, respectively, for each element. Relative
standard deviations (RSD) for Si, Na, and K were less than
1%, and others were less than 0.6%. Detection limits were
less than 500 ppm for all elements.
To aid cross-correlation of EPMA and ICPMS analyses
and to evaluate the general characteristics of crystal
123
Contrib Mineral Petrol (2011) 161:153–173
zoning, elemental distribution mapping of selected plagioclase crystals was performed with 15 kV, 20–50 nA,
and 1–3 lm for the acceleration voltage, beam current and
beam size, respectively. X-ray intensities of Ca-Ka, Na–
Ka, and K-Ka were counted for 0.02–0.03 s in each spot, at
1- to 3-lm intervals.
Comparison of EPMA-derived and LA-ICPMS-derived
anorthite contents
EPMA- and LA-ICPMS-derived anorthite contents are
compared for a number of plagioclase crystals in Fig. S1.
Due to the small EPMA beam size and closely spaced
sampling interval, a range in XAn is obtained over the
distance of 16–24 lm covered by the laser ablation spot
analysis (cf. Fig. S1c). In Fig. S1a, the average and the
respective range in EPMA-derived XAn is provided for
each LA-ICPMS data point. Data between the two dashed
lines are within the combined analytical uncertainties of
each technique. The results are largely consistent between
the two analytical methods, with exception of a few outliers, for which LA-ICPMS appears to slightly underestimate XAn. These outliers are more closely inspected in Fig.
S1b and S1c. During the laser ablation analysis of one
location from crystal W9604-C3-5, situated at the edge of a
calcic growth zone, there is a clear increase in the Na signal
with ablation time, suggesting that ablation progressed into
an underlying, more sodic growth zone (Fig. S1b). Another
example is crystal 2792-4R-2, where rim LA-ICPMS
analyses again appear to underestimate XAn. Here, the
outermost sodic rim contributes to the material sampled
during ablation but is not sampled during the spatially
restricted microprobe analysis (Fig. S1c).
The above comparison indicates that (i) LA-ICPMSderived anorthite contents that are based on plagioclase
stoichiometry are reliable, because they are generally
consistent with EPMA-derived anorthite contents, and (ii)
in cases where direct comparison of anorthite content with
trace element concentrations from the same spot is
required (as in this study), the use of LA-ICPMS-derived
XAn values is preferable, because they are based on
exactly the same volume of ablated material, such that
artefacts introduced by compositional heterogeneities
within that volume affect both major and trace elements
to a similar degree. We therefore use the LA-ICPMSderived XAn values here.
References
Allègre CJ, Condomines M (1976) Fine chronology of volcanic
processes using 238U–230Th systematics. Earth Planet Sci Lett
28:395–406
Contrib Mineral Petrol (2011) 161:153–173
Almeev R, Holtz F, Koepke J, Haase K, Devey C (2008) Depths of
partial crystallization of H2O-bearing MORB: phase equilibria
simulations of basalts at the MAR near Ascension Island (7–
11°S). J Petrol 49:25–45
Bergmanis EC, Sinton JM, Rubin KH (2007) Recent eruptive history
and magma reservoir dynamics on the southern East Pacific Rise
at 17°300 S. Geochem Geophys Geosyst 8:Q12O06
Bezos A, Humler E (2005) The Fe3?/RFe ratios of MORB glasses and
their implications for mantle melting. Geochim Cosmochim Acta
69:711–725
Blundy JD, Wood BJ (1991) Crystal-chemical controls on the
partitioning of Sr and Ba between plagioclase feldspar, silicate
melts, and hydrothermal solutions. Geochim Cosmochim Acta
55:193–209
Boudier F, Nicolas A, Ildefonse B (1996) Magma chambers in the
Oman ophiolite: fed from the top and the bottom. Earth Planet
Sci Lett 144:239–250
Canales JP, Detrick RS, Carbotte SM, Kent GM, Diebold JB, Harding
A, Babcock J, Nedimovic MR, van Ark E (2005) Upper crustal
structure and axial topography at intermediate spreading ridges:
seismic constraints from the southern Juan de Fuca ridge. J
Geophys Res 110:B12104
Carbotte SM, Detrick RS, Harding A, Canales JP, Babcock J, Kent G,
Van Ark E, Nedimovic M, Diebold J (2006) Rift topography
linked to magmatism at the intermediate spreading Juan de Fuca
Ridge. Geology 34:209–212
Cashman KV (1993) Relationship between plagioclase crystallization
and cooling rate in basaltic melts. Contrib Mineral Petrol
113(1):126–142
Cashman KV, Marsh BD (1988) Crystal size distribution (CSD) in
rocks and the kinetics and dynamics of crystallisation. II:
Makaopuhi lava lake. Contrib Mineral Petrol 99:292–305
Chadwick WW Jr, Embley RW, Fox CG (1995) SeaBeam depth
changes associated with recent lava flows, CoAxial segment,
Juan de Fuca Ridge: evidence for multiple eruptions between
1981–1993. Geophys Res Lett 22:167–170
Chadwick WW Jr, Embley RW, Shank TM (1998) The 1996 Gorda
Ridge eruption: geologic mapping, sidescan sonar, and SeaBeam
comparison results. Deap Sea Res II 45:2547–2567
Chen Y, Morgan WJ (1990) Rift valley/no rift valley transition at
mid-ocean ridges. J Geophys Res 95:17583–17604
Cherniak DJ, Watson EB (1994) A study of strontium diffusion in
plagioclase using Rutherford backscattering spectroscopy. Geochim Cosmochim Acta 58(23):5179–5190
Christie DM, Carmichael ISE, Langmuir CH (1986) Oxidation states
in mid-ocean ridge basalt glasses. Earth Planet Sci Lett 79:397–
411
Condomines M (1997) Dating recent volcanic rocks through
230
Th-238U disequilibrium in accessory minerals: example of
the Puy de Dome (French Massif Central). Geology 25(4):375–
378
Condomines M, Morand P, Allègre CJ (1981) 230Th-238U radioactive
disequilibria in tholeiites from the FAMOUS zone (Mid-Atlantic
Ridge, 36 deg 50 min N): Th and Sr isotopic geochemistry.
Earth Planet Sci Lett 55:247–256
Cooper KM, Goldstein SJ, Sims KWW, Murrell MT (2003) Uraniumseries chronology of Gorda Ridge volcanism: new evidence from
the 1996 eruption. Earth Planet Sci Lett 206:459–475
Costa F, Dungan M (2005) Short time scales of magmatic assimilation from diffusion modeling of multiple elements in olivine.
Geology 33:837–840
Costa F, Chakraborty S, Dohmen R (2003) Diffusion coupling
between trace and major elements and a model for calculation of
magma residence times using plagioclase. Geochim Cosmochim
Acta 67:2189–2200
171
Costa F, Coogan LA, Chakraborty S (2010) The time scales of mixing
and mingling involving primitive melts and melt-mush interaction at Mid-Ocean Ridges. Contrib Mineral Petrol 159:371–387
Davis AS, Clague DA (1987) Geochemistry, mineralogy, and
petrogenesis of basalt from the Gorda Ridge. J Geophys Res
92(B10):10467–10483
Davis AS, Clague DA (1990) Gabbroic xenoliths from the northern
Gorda Ridge: Implications for magma chamber processes under
slow spreading centers. J Geophys Res 95(B7):10885–10905
Davis AS, Clague DA, Cousens BL, Keaten R, Paduan JB (2008)
Geochemistry of basalt from the North Gorda segment of the
Gorda Ridge: evolution toward ultraslow spreading ridge lavas
due to decreasing magma supply. Geochem Geophys Geosyst
9:Q04004
DeMets C, Gordon RG, Argus DF, Stein S (1994) Effect of recent
revisions to the geomagnetic reversal time scale on estimates of
current plate motions. Geophys Res Lett 21:2191–2194
Dick HJB, Lin J, Schouten H (2003) An ultraslow-spreading class of
ocean ridge. Nature 426:405–412
Dowty E (1980) Synneusis reconsidered. Contrib Mineral Petrol
74:75–84
Duchene S, Pupier E, Le Carlier De Veslud C, Toplis MJ (2008) A 3D
reconstruction of plagioclase crystals in a synthetic basalt. Am
Mineral 93:893–901
Elthon D (1979) High magnesia liquids as the parental magma for
ocean floor basalts. Nature 278:514–518
Embley RW, Chadwick WW Jr, Jonasson IR, Butterfield DA, Baker
ET (1995) Initial results of the rapid response to the 1993
CoAxial event: relationships between hydrothermal and volcanic
processes. Geophys Res Lett 22:143–146
Embley RW, Chadwick WW Jr, Perfit MR, Smith MC, Delaney P
(2000) Recent eruptions on the CoAxial segment of the Juan de
Fuca Ridge: implications for mid-ocean ridge accretion processes. J Geophys Res 105(B7):16501–16525
Ernewein M, Pflumio C, Whitechurch H (1988) The death of an
accrecion zone as evidenced by the magmatic history of the
Sumail ophiolite (Oman). Tectonophys 151:247–274
Fox CG, Dziak RP (1998) Hydroacoustic detection of volcanic
activity on the Gorda Ridge, February-March 1996. Deep Sea
Res II 45:2513–2530
Fox CG, Radford WE, Dziak RP, Lau T-K, Matsumoto H, Schreiner
AE (1995) Acoustic detection of a seafloor spreading episode on
the Juan de Fuca Ridge using military hydrophone arrays.
Geophys Res Lett 22:131–134
Gerlach DC, Grove TL (1982) Petrology of Medicine Lake Highland
volcanics: characterisation of endmembers of magma mixing.
Contrib Mineral Petrol 80:147–159
Ghiorso MS, Sack RO (1995) 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
Giletti BJ, Casserly JED (1994) Strontium diffusion kinetics in
plagioclase feldspars. Geochim Cosmochim Acta 58(18):3785–
3793
Grove TL (1990) Cooling histories of lavas from Serocki Volcano.
In: Detrick R, Honnorez J, Bryan WB, Juteau T (eds) Proceedings of the ocean drilling program, Scientific Results, vol 106/
109, pp 3–8
Grove TL, Kinzler RJ, Bryan WB (1992) Fractionation of Mid-Ocean
Ridge Basalt (MORB). Mantle flow and melt generation at midocean ridges. In: Phipps Morgan J, Blackman DK, Sinton J (eds)
Geophysical monograph, vol 71, pp 281–310
Heath E, Turner S, Macdonald R, Hawkesworth CJ, van Calsteren P
(1998) Long magma residence times at an island arc volcano
123
172
(Soufriere, St. Vincent) in the Lesser Antilles: evidence from
238
U–230Th isochron dating. Earth Planet Sci Lett 160:49–63
Herzberg C (2004) Partial crystallization of mid-ocean ridge basalts
in the crust and mantle. J Petrol 45:2389–2405
Higgins MD (1996) Magma dynamics beneath Kameni volcano,
Thera, Greece, as revealed by crystal size and shape measurements. J Volcanol Geotherm Res 70:37–48
Higgins MD (2000) Measurement of crystal size distributions. Am
Mineral 85:1105–1116
Humler E, Whitechurch H (1988) Petrology of basalts from the
Central Indian Ridge (lat. 25°230 S, long. 70°040 E): estimates of
frequencies and fractional volumes of magma injections in a
two-layered reservoir. Earth Planet Sci Lett 88:169–181
Huppert HE, Sparks RSJ (1980) The fluid dynamics of a basaltic
magma chamber replenished by influx of hot dense ultrabasic
magma. Contrib Mineral Petrol 75:279–289
Kawamura K, Hosono T, Allawati HM, GOgawa Y, Taniguchi H
(2005) Origin of layering in cumulate gabbros in the Oman
ophiolite: Insights from magnetic susceptibility measurements in
the Wadi Sadm area. The Island Arc 14:564–570
Kelemen PB, Aharonov E (1998) Periodic formation of magma
fractures and generation of layered gabbros in the lower crust
beneath oceanic spreading ridges. In: Buck WR, Delaney PT,
Karson JA, Lagabrielle Y (eds) Faulting and magmatism at midocean ridges, Geophysical Monograph, vol 106. American
Geophysical Union, Washington, pp 267–289
Kelemen PB, Hirth G, Shimizu N, Spiegelman M, Dick HJB (1997a)
A review of melt migration processes in the adiabatically
upwelling mantle beneath oceanic spreading ridges. Phil Trans R
Soc A 355:283–318
Kelemen PB, Koga K, Shimizu N (1997b) Geochemistry of gabbro
sills in the crust-mantle transition zone of the Oman ophiolite:
implications for the origin of the oceanic lower crust. Earth
Planet Sci Lett 146:475–488
Klein EM, Langmuir CH (1987) Global correlations of ocean ridge
basalt chemistry with axial depth and crustal thickness. J
Geophys Res 92:8089–8115
Kohn SC, Henderson CMB, Mason RA (1989) Element zoning trends
in olivine phenocrysts from a supposed primary high-magnesian
andesite: an electron- and ion-microprobe study. Contrib Mineral
Petrol 103:242–252
Komar PD (1972) Mechanical interactions of phenocrysts and flow
differentiation of igneous dikes and sills. Geol Soc Am Bull
83:973–988
Langmuir CH (1989) Geochemical consequences of in situ crystallization. Nature 340:199–205
Langmuir CH, Bender JF, Batiza R (1986) Petrological and tectonic
segmentation of the East Pacific Rise, 5°300 –14°300 N. Nature
322:422–429
Langmuir CH, Klein EM, Plank T (1992) Petrological systematics of
mid-ocean ridge basalts: constraints on melt generation beneath
ocean ridges. In: Phipps Morgan J, Blackman DK, Sinton JM
(eds) Mantle flow and melt generation at mid-ocean Ridges.
Geophysical Monograph, vol 71, pp 183–280
Lejeune A, Richet P (1995) Rheology of crystal-bearing silicate
melts: an experimental study at high viscosity. J Geophys Res
100:4215–4229
Longerich HP, Jackson SE, Günther D (1996) Laser ablationinductively coupled plasma-mass spectrometric transient signal
data acquisition and analyte concentration calculation. J Anal At
Spectrom 11:899–904
Marsh BD (1988) Crystal size distribution (CSD) in rocks and the
kinetics and dynamics of crystallisation. I: theory. Contrib
Mineral Petrol 99:277–291
Martin D, Nokes R (1989) A fluid-dynamical study of crystal settling
in convecting magmas. J Petrol 30:1471–1500
123
Contrib Mineral Petrol (2011) 161:153–173
Menke W, West M, Tolstoy M (2002) Shallow-crustal magma
chamber beneath the axial high of the coaxial segment of Juan de
Fuca Ridge at the source site of the 1993 eruption. Geology
2002:359–362
Michael PJ, Cornell WC (1998) Influence of spreading rate and
magma supply on crystallization and assimilation beneath midocean ridges: evidence from chlorine and major element
chemistry of mid-ocean ridge basalts. J Geophys Res 103(B8):
18325–18356
Michael PJ, Langmuir CH, Dick HJB, Snow JE, Goldstein SL,
Graham DW, Lehnert K, Kurras G, Jokat W, Mühe R, Edmonds
HN (2003) Magmatic and amagmatic seafloor generation at the
ultraslow-spreading Gakkel ridge, Arctic Ocean. Nature
423:956–961
Morgan DJ, Blake S (2006) Magmatic residence times of zoned
phenocrysts: introduction and application of the binary element
diffusion modelling (BEDM) technique. Contrib Mineral Petrol
151:58–70
Morgan DJ, Blake S, Rogers NW, De Vivo B, Rolandi G, Macdonald
R, Hawkesworth CJ (2004) Timescales of crystal residence,
magma chamber volume from modelling of diffusion profiles
in phenocrysts: Vesuvius 1944. Earth Planet Sci Lett 222:933–
946
Morgan DJ, Jerram DA, Chertkoff DG, Davidson JP, Kronz A,
Nowell GM (2007) Combining CSD and isotopic microanalysis:
magma supply and mixing processes at Stromboli Volcano,
Aeolian Islands, Italy. Earth Planet Sci Lett 260:419–431
Morse SA (1984) Cation diffusion in plagioclase feldspar. Science
225:504–505
Nabelek PI, Langmuir CH (1986) The significance of unusual zoning
in olivines from FAMOUS area basalt 527–1-1. Contrib Mineral
Petrol 93:1–8
Nakamura M (1995) Continuous mixing of crystal mush and
replenished magma in the ongoing Unzen eruption. Geology
23(9):807–810
Niu Y, Batiza R (1991) An empirical method for calculating melt
compositions produced beneath mid-ocean ridges: application
for axis and off-axis (seamounts) melting. J Geophys Res
96:21753–21777
Niu Y, O’Hara MJ (2008) Global correlations of ocean ridge basalt
chemistry with axial depth: a new perspective. J Petrol 49:633–
664
Pan Y, Batiza R (2002) Mid-ocean ridge magma chamber processes:
Constraints from olivine zonation in lavas from the East Pacific
Rise at 9°300 N and 10°300 N. J Geophys Res 107:2022–2035
Perfit MR, Chadwick WW Jr (1998) Magmatism at mid-ocean ridges:
constraints from volcanological and geochemical investigations.
Geophys Monogr 106:59–115
Perfit MR, Fornari DJ, Smith MC, Bender JF, Langmuir CH, Haymon
RM (1994) Small-scale spatial and temporal variations in midocean ridge magmatic processes. Geology 22:375–379
Phipps Morgan J, Chen YJ (1993) Dependence of ridge-axis
morphology on magma supply and spreading rate. Nature
364:706–708
Pupier E, Duchene S, Toplis MJ (2008) Experimental quantification
of plagioclase crystal size distribution during cooling of a
basaltic liquid. Contrib Mineral Petrol 155:555–570
Pyle DM, Ivanovich M, Sparks RSJ (1988) Magma cumulate mixing
identified by U-Th disequilibrium dating. Nature 331:157–159
Qin R, Buck WR (2008) Why meter-wide dikes at oceanic spreading
centers? Earth Planet Sci Lett 265:466–474
Reynolds JR, Langmuir CH, Bender JF, Kastens KA, Ryan WBF
(1992) Spatial and temporal variability geochemistry of basalts
from the East Pacific Rise. Nature 359:493–499
Ridley WI, Perfit MR, Smith MC, Fornari DJ (2006) Magmatic
processes in developing oceanic crust revealed in a cumulate
Contrib Mineral Petrol (2011) 161:153–173
xenolith collected at the East Pacific Rise, 9°500 N. Geochem
Geophys Geosyst 7:Q12O04
Rubin AM (1995) Propagation of magma-filled cracks. Annu Rev
Earth Planet Sci 23:287–336
Rubin KH, Macdougall JD (1988) 226Ra excesses in mid-ocean-ridge
basalts and mantle melting. Nature 335:158–161
Rubin KH, Sinton JM (2007) Inferences on mid-ocean ridge thermal
and magmatic structure from MORB compositions. Earth Planet
Sci Lett 260:257–276
Rubin KH, Zellmer GF (2009) Reply to comment on ‘‘On the recent
bimodal magmatic processes and their rates in the TorfajökullVeidivötn area, Iceland’’ by K.M. Cooper. Earth Planet Sci Lett
281:115–123
Rubin KH, Smith MC, Perfit MR, Christie DM, Sacks LF (1998)
Geochronology and geochemistry of lavas from the 1996 North
Gorda Ridge eruption. Deep Sea Res II 45:2571–2597
Rubin KH, Smith MC, Bergmanis EC, Perfit MR, Sinton JM, Batiza
R (2001) Geochemical heterogeneity within mid-ocean ridge
lava flows: insights into eruption, emplacement and global
variations in magma generation. Earth Planet Sci Lett 2001:349–
367
Rubin KH, van der Zander I, Smith MC, Bergmanis EC (2005)
Minimum speed limit for ocean ridge magmatism from
210
Pb-226Ra-230Th disequilibria. Nature 437:534–538
Rubin KH, Sinton JM, Maclennan J, Hellebrandt E (2009) Magmatic
filtering of mantle composition at mid-ocean ridge volcanoes.
Nature Geosci 2:321–328
Schaefer SJ, Sturchio NC, Murrell MT, Williams SN (1993) Internal
238
U-series systematics of pumice from the November 13, 1985,
eruption of Nevado del Ruiz, Colombia. Geochim Cosmochim
Acta 57:1215–1219
Sims KWW, Goldstein SJ, Blichert-Toft J, Perfit MR, Kelemen P,
Fornari DJ, Michael P, Murrell MT, Hart SR, DePaolo DJ, Layne
G, Ball L, Jull M, Bender J (2002) Chemical and isotopic
constraints on the generation and transport of magma beneath the
East Pacific Rise. Geochim Cosmochim Acta 66:3481–3504
Sinton JM, Detrick RS (1992) Mid-ocean ridge magma chambers. J
Geophys Res 97:197–216
Sinton JM, Wilson DS, Christie DM, Hey RN, Delaney JR (1983)
Petrologic consequences of rift propagation on oceanic spreading
ridges. Earth Planet Sci Lett 62:193–207
Sinton JM, Smaglik SM, Mahoney JJ, Macdonald KC (1991)
Magmatic processes at superfast spreading mid-ocean ridges:
glass compositional variations along the East Pacific Rise. J
Geophys Res 96:6133–6155
Sinton J, Bergmanis E, Rubin K, Batiza R, Gregg TKP, Grönvold K,
Macdonald KC, White SM (2002) Volcanic eruptions on midocean ridges: New evidence from the superfast spreading East
Pacific Rise, 17°–19°S. J Geophys Res 107(B6):2115
Smith MC (1999) Geochemistry of eastern Pacific MORB: implications for MORB petrogenesis and the nature of crustal accretion
within the neovolcanic zone of two recently active ridge
segments. PhD thesis, University of Florida, Gainsville, 164 pp
Smith PM, Asimow PD (2005) Adiabat_1ph: A new public front-end
to the MELTS, pMELTS, and pHMELTS models. Geochem
Geophys Geosyst 6:Q02004
Smith MC, Perfit MR, Jonasson IR (1994) Petrology and geochemistry of basalts from the southern Juan de Fuca Ridge; controls
on the spatial and temporal evolution of mid-ocean ridge basalt.
J Geophys Res B99:4787–4812
173
Soule SA, Fornari DJ, Perfit MR, Rubin KH (2007) New insights into
mid-ocean ridge volcanic processes from the 2005–06 eruption
of the East Pacific Rise, 9°460 –560 N. Geology 35:1079–1082
Spence DA, Sharp PW, Turcotte DL (1987) Buoyancy-driven crack
propagation: a mechanism for magma migration. J Fluid Mech
174:135–153
Stolper E (1980) A phase diagram for mid-ocean ridge basalts.
Contrib Mineral Petrol 74:13–27
Sun SS, McDonough WF (1989) Chemical and isotopic systematics
of oceanic basalts: implications for mantle composition and
processes. In: Saunders AD, Norry MJ (eds) Magmatism in
ocean basins. Geological Society, London, Special Publications,
vol 42, pp 313–345
Thompson GW, Bryan B, Ballard R, Hamuro K, Melson WG (1985)
Axial processes along a segment of the East Pacific Rise, 10–
12 N. Nature 318:429–433
Turner SP, George RMM, Jerram DA, Carpenter N, Hawkesworth CJ
(2003) Case studies of plagioclase growth and residence times in
island arc lavas from Tonga and the Lesser Antilles, and a model
to reconcile discordant age information. Earth Planet Sci Lett
214:279–294
van Achterbergh E, Ryan CG, Jackson SE, Griffin WL (2001) Data
reduction software for LA-ICP-MS: Appendix. In: Sylvester PJ
(ed) Laser Ablation-ICP-mass spectrometry in the earth sciences:
principles and applications. Mineralogical Association of Canada Short Course Series, vol 29, pp 239–243
Vance JA (1969) On synneusis. Contrib Mineral Petrol 24:7–29
Vance JA, Gilreath JP (1967) The effect of synneusis on phenocryst
distribution patters in some porphyritic igneous rocks. Am
Mineral 52:529–536
Van Tongeren JA, Kelemen PB, Hanghøj K (2008) Cooling rates in
the lower crust of the Oman ophiolite: Ca in olivine, revisited.
Earth Planet Sci Lett 267:69–82
Volpe AM, Hammond PE (1991) 238U–230Th-226Ra disequilibria in
young Mount St. Helens rocks: time constraint for magma
formation and crystallisation. Earth Planet Sci Lett 107:475–486
Zellmer GF, Annen C (2008) An introduction to magma dynamics.
In: Annen C, Zellmer GF (eds) Dynamics of crustal magma
transfer, storage and differentiation. Geological Society, London,
Special Publications, vol 304, pp 1–13
Zellmer GF, Clavero J (2006) Using trace element correlation patterns
to decipher a sanidine crystal growth chronology: an example
from Taapaca volcano, Central Andes. J Volcanol Geotherm Res
156:291–301
Zellmer GF, Blake S, Vance D, Hawkesworth C, Turner S (1999)
Plagioclase residence times at two island arc volcanoes (Kameni
islands, Santorini, and Soufriere, St. Vincent) determined by Sr
diffusion systematics. Contrib Mineral Petrol 136(4):345–357
Zellmer GF, Turner SP, Hawkesworth CJ (2000) Timescales of
destructive plate margin magmatism: new insights from Santorini, Aegean volcanic arc. Earth Planet Sci Lett 174(3–4):265–
282
Zellmer GF, Sparks RSJ, Hawkesworth CJ, Wiedenbeck M (2003)
Magma emplacement and remobilization timescales beneath
Montserrat: insights from Sr and Ba zonation in plagioclase
phenocrysts. J Petrol 44:1413–1431
Zellmer GF, Rubin KH, Grönvold K, Jurado-Chichay Z (2008) On the
recent bimodal magmatic processes and their rates in the
Torfajökull-Veidivötn area, Iceland. Earth Planet Sci Lett
269:387–397
123
Contrib Mineral Petrol (2011) 161:175–176
DOI 10.1007/s00410-010-0595-y
ERRATUM
Erratum to: Crystal growth during dike injection of MOR basaltic
melts: evidence from preservation of local Sr disequilibria
in plagioclase
Georg F. Zellmer • Kenneth H. Rubin •
Peter Dulski • Yoshiyuki Iizuka •
Steven L. Goldstein • Michael R. Perfit
Published online: 27 November 2010
Ó Springer-Verlag 2010
Erratum to: Contrib Mineral Petrol
DOI 10.1007/s00410-010-0518-y
In the original paper, repeat analyses of gem labradorite
AMNH 95557 provided the basis for assessment of
relative uncertainties for individual elements and these
were cited as 1.5% for XAn, 1.2% for Sr and 2.6% for
Ba (1r, n = 16). Unfortunately, the beam conditions used
during the ablation of this standard were not identical to
The online version of the original article can be found under
doi:10.1007/s00410-010-0518-y.
G. F. Zellmer (&) Y. Iizuka
Institute of Earth Sciences, Academia Sinica, 128 Academia
Road, Section 2, Nankang, Taipei 11529, Taiwan, ROC
e-mail: [email protected]
G. F. Zellmer S. L. Goldstein
Lamont-Doherty Earth Observatory of Columbia University,
61 Route 9W, Palisades, NY 10964, USA
K. H. Rubin
Department of Geology and Geophysics, SOEST,
University of Hawaii at Manoa, 1680 East West Road,
Honolulu, HI 96822, USA
P. Dulski
Section 3.3, GFZ German Research Centre for Geosciences,
Helmholtz Centre Potsdam, Telegrafenberg,
14473 Potsdam, Germany
M. R. Perfit
Department of Geological Sciences,
University of Florida, Gainsville, FL 32611, USA
those used during sample analysis. We have recently
reassessed the analytical uncertainties, employing the
same analytical conditions as for our sample runs and
using a homogeneous growth zone of one of the crystal
samples as a working standard. We found that the relative uncertainties were in fact 0.5% for XAn, 1.9% for Sr
and 15% for Ba (1r, n = 30). Thus, analytical uncertainties stated in the original paper were overestimated
for XAn, underestimated for Sr and significantly underestimated for Ba.
As a result, the precision on Ba is in fact often too low to
resolve potential local Ba disequilibria. Many crystals are
within error of complete chemical equilibrium for Ba. The
conclusions of our study, however, were based on the
observed local Sr disequilibria. Using the revised uncertainties in our calculations, we find that two of the 23
crystals studied (JdF-2794-2R-4 and Gakkel D27-16-O) are
in fact within error of complete chemical equilibrium for
Sr. Further, three other crystals (JdF-2794-2R-1, JdF-27942R-7 and Gorda W9604-C3-3) have growth zones of
slightly more than 200 lm in width that are within error of
local Sr equilibrium, but do preserve local Sr disequilibria.
In the other crystals, the width of zones within local Sr
equilibrium rarely exceeds 100 lm, as stated in our original
contribution.
Below, we provide the corrected version of Table 2,
which summarizes our results. The great majority of
crystals preserves local Sr disequilibria and displays only
narrow zones in local Sr equilibrium. Equilibration times
remain of the order of months to a few years, and minimum
crystal growth rates remain of the order of 10-9–10-11 cm
s-1. The original conclusions of our paper with regard to
crystal residence times, growth rates and processes, and the
typical size of melt lenses within the gabbroic rocks in
oceanic layer 3, thus remain valid.
123
176
Contrib Mineral Petrol (2011) 161:175–176
Table 2 Determining minimum growth rates for all studied crystals
Crystal
XAn range
Min. T (°C)a
ÐSr (cf. Fig. 6)
(10-17 m2 s-1)
JdF 2792-4R-1
0.64–0.72
1,080
*1
JdF 2792-4R-2
0.67–0.79
1,080
*1
94
JdF 2792-4R-4
0.62–0.72
1,080
*1
132
JdF 2792-4R-5
0.67–0.79
1,080
*1
27
JdF 2792-4R-6
0.68–0.77
1,080
*1
JdF 2792-4R-7
0.69–0.74
1,080
JdF 2794-2R-1
JdF 2794-2R-2
0.65–0.71
0.66–0.68
JdF 2794-2R-3
JdF 2794-2R-4
Zone width
2b (lm)
Equil. time
b2/ÐSr (years)b
Distance
to rim (lm)
Minimum growth rate
(10-10 cm s-1)c
0.2
117
16.2
7.0
79
0.4
91
0.2
0.6
85
4.7
26
0.5
94
5.6
*1
28
0.6
68
3.5
1,100
1,100
*1
*1
44
29
1.5
0.7
106
78
2.2
3.7
0.67–0.72
1,100
*1
25
0.5
102
6.5
0.67–0.69
1,100
*1
JdF 2794-2R-5
0.67–0.72
1,100
*1
30
0.7
JdF 2794-2R-6
0.67–0.70
1,100
*1
55
2.4
87
1.2
JdF 2794-2R-7
0.67–0.71
1,100
*1
47
1.7
110
2.0
Gorda W9604-C3-1
0.79–0.82
1,130
*1
69
3.8
61
0.5
Gorda W9604-C3-2
0.75–0.80
1,130
*1
56
2.5
90
1.1
Gorda W9604-C3-3
0.73–0.76
1,130
*1
24
0.5
12
0.8
Gorda W9604-C3-4
0.70–0.83
1,130
*1
49
1.9
126
2.1
Gorda W9604-C3-5
0.70–0.72
1,130
*1
26
0.5
61
3.6
Gorda W9604-C3-6
0.71–0.81
1,130
*1
34
0.9
107
3.7
Gorda W9604-C3-7
0.76–0.84
1,130
*1
32
0.8
147
5.7
Gakkel D27-16-E
0.81–0.84
1,145
*1
105
8.7
683
2.5
Gakkel D27-16-H
Gakkel D27-16-O
0.79–0.84
0.81–0.84
1,145
1,145
*1
*1
87
6.0
832
4.4
Within error of equilibrium, no information retained
17
14
Within error of equilibrium, no information retained
43
1.9
a
Based on lowest anorthite content of all crystals in each rock sample and assuming 0.5 wt% H2O in the primitive magma
b
Actual crystal residence times are likely at least one order of magnitude lower, see text for discussion
c
These are effective minimum growth rates that do not account for limited intermittent dissolution. Actual growth rates are likely at least one
order of magnitude greater, see text for discussion
123