ChapterPDF Available

Halogens in Seawater, Marine Sediments and the Altered Oceanic Lithosphere

Authors:

Abstract and Figures

This chapter aims to provide a framework for understanding the distribution of halogens in the oceanic lithosphere. It reviews the concentrations of F, Cl, Br and I in seawater, marine sediment pore waters, hydrothermal vent fluids, fluid inclusions from deeper in the crust, and the complementary solid-phase reservoirs of organic matter and minerals present in sediments and crustal/mantle rocks from varying depths. Seawater (3.4–3.5 wt% salt) is depleted in F, weakly enriched in I and strongly enriched in Cl and Br compared to the primitive mantle. Sequestration of I and Br by phytoplankton lead to the storage of these elements in marine sediments, which are the Earths dominant I reservoir. Regeneration of organic matter during diagenesis releases I⁻ and Br⁻ to marine sediment pore waters, which acquire Br/Cl and I/Cl ratios of higher than seawater and can be advected into the underlying crust and lithosphere. In contrast, Cl is usually assumed to behave conservatively in pore waters and F is precipitated in authigenic sedimentary minerals meaning it is not significantly advected into the underlying basement. Vent fluids have salinities of 0.1–6 wt% salts, which provide evidence for phase separation and segregation of vapours and brines in hydrothermal systems. The majority of vent fluids have Br/Cl ratios within 10% of the seawater value. However, elevated Br/Cl and I/Cl ratios indicate that some vent fluids interact with organic-rich sediments, and low Br/Cl ratios suggest some vent fluids leach Cl from glassy volcanic rocks or halite. Vent fluids have F/Cl ratios scattered around the seawater value which reflects the generally low mobility of F during diagenesis and hydrothermal alteration. In comparison to vent fluids, fluid inclusions also provide evidence for phase separation but preserve a much greater range of salinity including brines with salinities as high as ~50 wt% salt. The altered ocean crust has a F concentration of close to its initial value. In contrast, Cl is mobilised within layer 2 pillows and dykes and strongly enriched in layer 3 gabbros subjected to high temperature alteration. Amphibole is the dominant Cl host in the oceanic crust, with Cl concentrations of <500 ppm under greenschist conditions and up to wt% levels under amphibolite conditions. The increasing Cl content of amphibole as a function of metamorphic grade most likely reflects a decreasing water/rock ratio and a general increase in fluid salinity as a function of depth in the crust. Amphibole preferentially incorporates Cl relative to Br and I; however, I is enriched in absolute terms, and relative to Cl, in clay-rich alteration and biogenic alteration of glassy rocks in the upper crust. Serpentinites formed in the oceanic lithosphere can contain thousands of ppm Cl and some serpentinites preserve Br/Cl and I/Cl signatures very similar to sedimentary pore waters, indicating that all halogens have high compatibilities in serpentine. Fluorine is slightly enriched in some serpentinites compared to peridotites, which may indicate minor mobilisation of F from igneous lithologies in the overlying crust. The altered oceanic crust and mantle lithosphere reaching subduction zones have poorly defined halogen concentrations. However, the average Cl concentration could be as high as ~400 ppm. And it may have a F/Cl ratios as low as ~0.25 compared to ~2 in pristine crust. It is estimated that approximately 90% of the Cl present in altered oceanic lithosphere is introduced during seawater alteration.
Content may be subject to copyright.
Chapter 9
Halogens in Seawater, Marine Sediments
and the Altered Oceanic Lithosphere
Mark A. Kendrick
Abstract This chapter aims to provide a framework for understanding the distri-
bution of halogens in the oceanic lithosphere. It reviews the concentrations of F, Cl, Br
and I in seawater, marine sediment pore waters, hydrothermal vent uids, uid
inclusions from deeper in the crust, and the complementary solid-phase reservoirs of
organic matter and minerals present in sediments and crustal/mantle rocks from
varying depths. Seawater (3.43.5 wt% salt) is depleted in F, weakly enriched in I and
strongly enriched in Cl and Br compared to the primitive mantle. Sequestration of I
and Br by phytoplankton lead to the storage of these elements in marine sediments,
which are the Earths dominant I reservoir. Regeneration of organic matter during
diagenesis releases I
and Br
to marine sediment pore waters, which acquire Br/Cl
and I/Cl ratios of higher than seawater and can be advected into the underlying crust
and lithosphere. In contrast, Cl is usually assumed to behave conservatively in pore
waters and F is precipitated in authigenic sedimentary minerals meaning it is not
signicantly advected into the underlying basement. Vent uids have salinities of
0.16 wt% salts, which provide evidence for phase separation and segregation of
vapours and brines in hydrothermal systems. The majority of vent uids have Br/Cl
ratios within 10% of the seawater value. However, elevated Br/Cl and I/Cl ratios
indicate that some vent uids interact with organic-rich sediments, and low Br/Cl
ratios suggest some vent uids leach Cl from glassy volcanic rocks or halite. Vent
uids have F/Cl ratios scattered around the seawater value which reects the generally
low mobility of F during diagenesis and hydrothermal alteration. In comparison to
vent uids, uid inclusions also provide evidence for phase separation but preserve a
much greater range of salinity including brines with salinities as high as *50 wt%
salt. The altered ocean crust has a F concentration of close to its initial value. In
contrast, Cl is mobilised within layer 2 pillows and dykes and strongly enriched in
layer 3 gabbros subjected to high temperature alteration. Amphibole is the dominant
Cl host in the oceanic crust, with Cl concentrations of <500 ppm under greenschist
conditions and up to wt% levels under amphibolite conditions. The increasing Cl
M.A. Kendrick (&)
Research School of Earth Sciences, The Australian National University, Canberra 2601,
ACT, Australia
e-mail: mark.kendrick@anu.edu.au
©Springer International Publishing AG 2018
D.E. Harlov and L. Aranovich (eds.), The Role of Halogens in Terrestrial
and Extraterrestrial Geochemical Processes, Springer Geochemistry,
https://doi.org/10.1007/978-3-319-61667-4_9
591
content of amphibole as a function of metamorphic grade most likely reects a
decreasing water/rock ratio and a general increase in uid salinity as a function of
depth in the crust. Amphibole preferentially incorporates Cl relative to Br and
I; however, I is enriched in absolute terms, and relative to Cl, in clay-rich alteration
and biogenic alteration of glassy rocks in the upper crust. Serpentinites formed in the
oceanic lithosphere can contain thousands of ppm Cl and some serpentinites preserve
Br/Cl and I/Cl signatures very similar to sedimentary pore waters, indicating that all
halogens have high compatibilities in serpentine. Fluorine is slightly enriched in
some serpentinites compared to peridotites, which may indicate minor mobilisation
of F from igneous lithologies in the overlying crust. The altered oceanic crust and
mantle lithosphere reaching subduction zones have poorly dened halogen con-
centrations. However, the average Cl concentration could be as high as *400 ppm.
And it may have a F/Cl ratios as low as *0.25 compared to *2 in pristine crust. It is
estimated that approximately 90% of the Cl present in altered oceanic lithosphere is
introduced during seawater alteration.
9.1 Introduction
As a chemical group, the halogens (F, Cl, Br, I) are united by the ability to form
negatively charged anions; however, Cl is uniquely important within this group
because it is the dominant ligand that enables metal transport in the majority of
hydrothermal solutions (Bischoff and Dickson 1975; Seyfried and Bischoff 1981;
Yardley 2005). Fluid salinity is consequently of profound importance to the mass
transfer processes operating during hydrothermal alteration of the oceanic litho-
sphere. There is wide interest in the scale of seaoor alteration processes and meta-
somatism because alteration of the oceanic lithosphere moderates the composition of
seawater (Bischoff and Dickson 1975; Edmond et al. 1979) and determines the
chemical composition of the slab that is subducted into the mantle at convergent plate
margins. Seaoor alteration is therefore critical to the geological cycles of all ele-
ments (Bischoff and Dickson 1975; Edmond et al. 1979; Spandler and Pirard 2013).
In comparison to Cl, the other halogens exhibit systematically different beha-
viours in the oceanic lithosphere that are related to their ionic sizes or elemental
abundances. Fluorine and Cl are similarly abundant on Earth with primitive mantle
concentrations on the order of *1530 ppm (Table 9.1; Kendrick et al. 2017; Pyle
and Mather 2009). However, the F
anion is signicantly smaller than the heavier
halogens and F is strongly electronegative, being the only element in the periodic
table that is more electronegative than oxygen. These properties result in F having a
much higher compatibility than Cl in many hydrous and nominally anhydrous
minerals (Bernini et al. 2013; Carpenter 1969; Dalou et al. 2012; Frohlich et al. 1983;
Ichikuni 1979; Seyfried and Ding 1995). Consequently, F is distinguished from the
heavier halogens by a low solubility in aqueous uids (Seyfried and Ding 1995).
Bromine and I have even higher solubilities in aqueous uids than Cl, but they
are trace constituents on Earth with primitive mantle concentrations of *76 ppb for
592 M.A. Kendrick
Br and *7 ppb for I (Table 9.1). Iodine is a weakly electronegative redox sensitive
element that is important in a number of biochemical pathways, which together with
its low abundance mean it is regarded as an essential element for life (Eldereld and
Truesdale 1980; Fuge and Johnson 1986; Leblanc et al. 2006).
The contrasting solubilities of F, Cl, Br and I in aqueous uids mean that their
relative abundances, or the F/Cl, Br/Cl and I/Cl, ratios of a uid, can provide useful
information about a number of hydrogeochemical processes relevant to alteration of
the oceanic lithosphere. These include the roles of phase separation, crustal-
hydration, dissolution of evaporites, or precipitation of hydrothermal-halite in
controlling vent uid salinity (Campbell and Edmond 1989; Oosting and Von
Damm 1996; Seyfried et al. 2003; You et al. 1994). In addition, the presence of
organic matter can be inferred from high I abundances (Campbell and Edmond
1989; Gieskes et al. 2002; Kawagucci et al. 2011). This information is of consid-
erable importance to investigating the role of uids in alteration of the oceanic
lithosphere and the depth of the biosphere (Campbell and Edmond 1989; Mottl
et al. 2011; Oosting and Von Damm 1996; Reeves et al. 2011; Wu et al. 2012; You
et al. 1994).
The current chapter aims to show how the different geochemical behaviours of
the halogens control their abundances in both hydrothermal uids and the com-
plementary reservoirs of the oceanic lithosphere and how salts, and acids generated
by hydrolysis of salts, facilitate transport of metals in solution (Seyfried and
Bischoff 1981). The majority of data for both altered oceanic rocks and uids are
for Cl, with more limited data available for F and Br, and very limited data for I.
Therefore while the aim is to provide a framework for understanding the distri-
bution of halogens in the oceanic lithosphere, an important role of this chapter is
also to identify current gaps in our knowledge. Improving constraints on the
absolute and relative abundances of F, Cl, Br, and I in different lithospheric
reservoirs (e.g., sediments, altered basalt, serpentinites) is critical for optimal use of
halogen abundance ratios in tracking halogens through the subduction cycle
(Kendrick et al. 2012b,2017; Straub and Layne 2003) and will further contribute to
our understanding of marine halogen cycling (Frohlich et al. 1983; Leblanc et al.
2006; Seyfried et al. 1986; Seyfried and Ding 1995).
Table 9.1 Halogens in
seawater and the primitive
mantle
F ppm Cl ppm Br ppb I ppb
Primitive
mantle
17 ±626±876±25 7 ±4
Seawater 1.3 19,350 67,300 58
Seawater/PM *0.08 *740 *890 *8
References: Drever (1997); Eldereld and Truesdale (1980);
Huang et al. (2005); primitive mantle concentrations of Kendrick
et al. (2017), see also McDonough and Sun (1995), Kamenetsky
and Eggins (2012) and Pyle and Mather (2009)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 593
9.2 Halogen Chemistry in Seawater and Marine
Sediments
The halogen composition of seawater and the more evolved uids that have
interacted with sediments and could be advected into the mac portions of the
oceanic lithosphere is controlled by the unique chemical properties of each halogen
and their abundances on Earth.
The concentrations of F, Cl and Br in seawater, present as the uoride (F
),
chloride (Cl
), and bromide (Br
) anions, is fairly constant over thousand-year
timescales with minor variations related to salinity which uctuates around *3.4
3.5 wt% salts (Drever 1997; Pinet 1992). In contrast, I is unique among the
halogens because its predominant species in seawater is iodate (IO
3
) and iodide
(I
) is only important in surface waters and small anoxic basins (Eldereld and
Truesdale 1980; Huang et al. 2005; Wong and Brewer 1977). The concentration of
iodate, which is taken up by a number of marine organisms, is correlated with other
nutrients such as nitrate and phosphate and varies as a function of water depth and
latitude (Eldereld and Truesdale 1980; Huang et al. 2005). In addition to iodate
and iodide, dissolved organic I is signicant at shallow and intermediate depths
accounting for up to *5 ppb of the total I at depths of 400 m in the north Pacic
(Huang et al. 2005). The total I concentration of seawater varies from about 44 ppb
up to 76 ppb with a mean of *58 ppb (Eldereld and Truesdale 1980; Huang et al.
2005; Nakayama et al. 1989; Zheng et al. 2012).
The average abundances of halogens in seawater (Drever 1997; Eldereld and
Truesdale 1980; Huang et al. 2005) and estimated abundances of halogens in the
primitive mantle (Kendrick et al. 2017; McDonough and Sun 1995; Palme and
ONeill 2003; Sharp and Draper 2013) are given in Table 9.1. These data illustrate
that F is unique among the halogens because it is depleted in seawater relative to the
primitive mantle (Table 9.1). In comparison, I exhibits a modest enrichment in
seawater and Cl and Br are both enriched by hundreds of times in seawater com-
pared to the primitive mantle (Table 9.1).
9.2.1 Fluorines Low Solubility
The low solubility of F in aqueous uids means that in contrast to the heavy
halogens (Cl, Br and I), F is inefciently transported by rivers, and is removed from
solution once in the ocean (Carpenter 1969). Important F sinks in the oceanic
environment include marine carbonates, phosphates, and alumina-silicates
(clay minerals) (cf., Carpenter 1969; Frohlich et al. 1983; Matthies and Troll
1990; Rude and Aller 1994). Calcite can contain hundreds of ppm of F and
magnesium-calcite and aragonite have maximum F concentrations in excess of
1000 ppm that represent a 1000-fold enrichment over seawater and demonstrate
the high compatibility of F in carbonate (Fig. 9.1; Ohde and Kitano 1980;
594 M.A. Kendrick
Okumura et al. 1983; Opdyke et al. 1993; Rude and Aller 1991; Tanaka et al.
2013). An exchange reaction has been proposed in which two F
anions substitute
for CO
3
2
(Ichikuni 1979):
CaCO3þ2F¼CaF2þCO2
3ð9:1Þ
Given the conservative nature of F
in seawater, this reaction indicates that the F
content of aragonite can be used to monitor the CO
2
content of seawater, which
exhibits signicant spatial and temporal variation (e.g., Ichikuni 1979; Ramos et al.
2005; Tanaka et al. 2013).
Carbonate uorapatite (CFA) in marine phosphates with concentrations of 13
wt% F contains an order of magnitude more F than carbonate (Li and Schoonmaker
2003; Rude and Aller 1991). However, because phosphorite deposits are restricted
to zones of nutrient rich water upwelling they are estimated to account for only
*20% of F uptake from the ocean (Frohlich et al. 1983). Fluorine also substitutes
for the hydroxyl group in alumina-silicates (clay minerals); however, the relative
importance of this sink is disputed (cf., Carpenter 1969; Matthies and Troll 1990;
Rude and Aller 1994).
Fluorine is mobilised during diagenesis of marine carbonates but it is xed into
newly formed carbonates, apatite, or clay minerals, meaning the uids occupying
sediment pore spaces have low ppm levels of F and F/Cl ratios similar to seawater
(Frohlich et al. 1983; Mahn and Gieskes 2001; Rude and Aller 1994).
Fluorine is further distinguished from the heavier halogens by its relative com-
patibility in the mantle. Fluorine has a compatibility comparable to moderately
incompatible elements like Nd or Pr (Kendrick et al. 2017; Workman et al. 2006),
whereas the heavy halogens (Cl, Br and I) have incompatibilities comparable to more
strongly incompatible elements such as K, La or Nb (Kendrick et al. 2012a,2017;
Fig. 9.1 The F content
(ppm) of different types of
carbonate. Mg carbonates and
aragonite have higher F
contents than calcite
(Carpenter 1969; Ohde and
Kitano 1980; Ramos et al.
2005; Rude and Aller 1991;
Tanaka et al. 2013)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 595
le Roux et al. 2006; Schilling et al. 1980). This difference cannot explain the F
depletion of seawater but it explains why the surface reservoirs of seawater and
sediments are less enriched in F than the heavy halogens (cf., Table 9.1; Kendrick
et al. 2017).
9.2.2 Chlorine in Halite
In contrast to F, Cl is usually regarded as a conservative element in sedimentary
marine pore waters, meaning that its concentration is not easily altered. Chlorine
and Br have very low compatibilities in marine carbonate (Kitano et al. 1975;
Okumura et al. 1986) and low concentrations in other sedimentary minerals such as
carbonate uorapatite. As a result, evaporites formed by evaporation of seawater
beyond the point of halite saturation (3035 wt% salt; McCaffrey et al. 1987;
Zherebtsova and Volkova 1966), commonly represent the only signicant mineral
reservoir of Cl in sedimentary settings (Hanor 1994; Worden 1996). Evaporite
deposits are important sedimentary units in some ocean basins and dissolution of
evaporites is indicated as an important source of salinity for hydrothermal uids in
Red Sea brine pools (e.g., Pierret et al. 2001; Shanks and Bischoff 1977; Zierenberg
and Shanks III 1986). In addition, halite could be important in some hydrothermal
settings where precipitation of hydrothermal halite is possible (Buttereld et al.
1997; Foustoukos and Seyfried 2007).
Evaporitic salt deposits are characterised by very low Br/Cl and I/Cl ratios (Bein
et al. 1991; Fontes and Matray 1993; McCaffrey et al. 1987). Halite has rela-
tive partition coefcients (D = [X/Cl]
salt
/[X/Cl]
uid
)of*0.03 for Br/Cl and <0.003
for I/Cl, whereas Br can be accommodated more easily in sylvite (D *0.2), carnalite,
and other potash salts (Fontes and Matray 1993; Hermann 1980; Holser 1979;
Siemann and Schramm 2000). As a result, the actual concentrations of Br and I in
evaporite deposits vary depending on the salts present; the presence of organic
material and uid inclusions, which contain signicant Br and I; and on the diage-
netic history of the deposit (Bein et al. 1991; Fontes and Matray 1993). Experimental
studies suggest that Br is excluded from the halite lattice under hydrothermal con-
ditions (Berndt and Seyfried 1997; Foustoukos and Seyfried 2007). This means that
uid interaction with either sedimentary or hydrothermal halite has the potential to
strongly alter Cl concentrations and uid Br/Cl and I/Cl ratios.
9.2.3 Iodine and Bromine in Organic Matter
The marine I cycle is dominated by accumulation of I in algae: Kelp (a type of
seaweed, which is multicellular macroalgae) is the main I accumulator in coastal
environments, whereas phytoplankton is the main I accumulator at sea (Iwamoto
and Shiraiwa 2012; Leblanc et al. 2006).
Iodine is seasonally variable and heterogeneously distributed in Kelp (Gall et al.
2004). Maximum concentrations of *6 wt% I have been reported in parts of a Kelp
596 M.A. Kendrick
plant (Laminaria Digitalis; Kuepper et al. 2013), but an average I content of 1 wt%
has been suggested (Gall et al. 2004). Other seaweeds contain *1500 ppm
I (Muramatsu and Wedepohl 1998; Yang et al. 2014). Several species of coral have
skeletons in which organic components can have very high halogen concentrations
of 56 wt% I, 12 wt% Br, and <1000 ppm Cl (Collins 1969; Goldberg et al. 1994).
The Br/I weight ratios of seaweeds and corals range from *0.1 to 2.5 (Collins 1969;
Kuepper et al. 2013). In comparison, plankton has variable halogen contents ranging
from sub-ppm to typical values of *200300 ppm I and 10004000 ppm Br.
Plankton have Br/I weight ratios of 324 and I/C
organic
of 0.00020.002 (Bobrov
et al. 2005; Iwamoto and Shiraiwa 2012; Martin et al. 1993; Price and Calvert 1977).
Iodine accumulated by algae is either emitted to the atmosphere as I
2
gas (or
volatile iodocarbon compounds) or buried in marine sediment, which has previously
been estimated to contain *70% of the I present in the crust and seawater (Leblanc
et al. 2006; Muramatsu and Wedepohl 1998). Regeneration of organic matter releases
iodide (I
), bromide (Br
), and CH
4
into sedimentary pore waters in the subsurface
(Kennedy and Eldereld 1987a,b; Muramatsu et al. 2007; Price and Calvert 1973,
1977). Studies of cosmosgenic
129
I indicate pore waters can migrate laterally over
many km (e.g., Fehn et al. 2006; Muramatsu et al. 2001,2007). However, pore waters
also seep gradually upwards toward the seaoor (Wong and Brewer 1977). In typical
oxic basins, iodide in escaping pore waters is oxidised and, together with seawater
iodate, is then adsorbed back onto organic detritus close to the sediment surface
(Fig. 9.2; Price and Calvert 1973,1977). If present, additional adsorption may also
occur on Fe-hydroxides (Ullman and Aller 1985) and clay minerals (Montavon et al.
2014). As a result sediment proles in oxic basins can have very high I concentrations
of 1002000 ppm close to the surface (Fig. 9.2a; Price and Calvert 1977), along with
elevated I/C
organic
and I/Br ratios, but at depths of more than a metre below the
surface, sediments have more typical concentrations of *550 ppm I, along with
I/C
organic
and I/Br ratios similar to plankton (Kennedy and Eldereld 1987a,b; Price
and Calvert 1973,1977; Upstillgoddard and Eldereld 1988).
Halogen regeneration processes result in pore water Br and I concentrations
tending to increase with depth and solid phase Br and I concentrations tending to
decrease with depth and sediment maturity (Fig. 9.2). Pelagic sediments have typical
I concentrations of *550 ppm I in the solid phase (Kennedy and Eldereld 1987b;
Muramatsu et al. 2007). However, Murumatsu et al. (2007) reported that sediments
from the Nankai Trough contain 34 times more Br and I in pore waters than in the
solid phase. In contrast to unconsolidated sediment, lithied marine sediments have
much lower halogen concentrations. For example, the median I concentration of
twenty limestones and shales analysed by Muramatsu and Wedepohl (1998)is2
þ2
1
ppm. The Callovian-Oxfordian clay of France contains 17 ppm I and 0.41.2 wt%
organic C (Claret et al. 2010). The organic rich Kimmeridge clay of England con-
tains 634 ppm I and 1655 wt% organic C (Cosgrove 1970). Taken together, these
data give a combined average of 4 ±3 ppm I in lithied marine sediments
(Table 9.2) and suggest that variable depletion of I occurs relative to C in lithied
sediments, compared with organic I/C ratios.
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 597
A large database is now available for Cl, Br and I in sedimentary marine pore
waters (Fig. 9.3; Fehn et al. 2000,2003,2006,2007a,b; Gieskes and Mahn 2007;
Mahn and Gieskes 2001; Martin et al. 1993; Muramatsu et al. 2001,2007; Tomaru
et al. 2007a,b,c,2009). These studies have shown that sedimentary pore waters
have a relatively narrow range in salinity (*15 wt% salt), which reects the
modication of seawater salinity (3.5 wt% salts) by formation and destruction of gas
hydrate (Fig. 9.3b). Plotting pore water Cl, Br, and I data as ratios in a three element
plot, which removes the effects of salinity variation, enables the sediment-derived
halogen component to be fully resolved from seawater (Fig. 9.3a). The slopes in
Fig. 9.3a indicate that the sediment-derived halogen component has a fairly constant
seawater-corrected Br/I weight ratio of *0.31.6 (Fig. 9.3a; Kendrick et al. 2011a).
Presumably the low seawater-corrected Br/I ratio of sedimentary pore waters
Fig. 9.2 Iodine in dry
sediments (solid phase) and
sedimentary pore waters as a
function of depth. aThe top
80 cm of sediments on the
Namibian Shelf that contain
28 wt% organic carbon are
exceptionally enriched in I
(Price and Calvert 1977), but
show a typical surface
enrichment of I resulting from
iodate adsorption (Kennedy
and Eldereld 1987a,b; Price
and Calvert 1977). bComplex
deep sediment proles on the
Nankai Trough show that
deeper pore uids have
generally higher I
concentrations than shallower
pore uids; this trend is seen
most clearly in the top 100 m
of the second prole
(Muramatsu et al. 2007)
598 M.A. Kendrick
(0.31.6), compared to plankton (324; Bobrov et al. 2005; Collins 1969; Martin
et al. 1993), reects the surface-adsorbed I component acquired on the
seaoor under oxic conditions (Francois 1987; Martin et al. 1993; Upstillgoddard
and Eldereld 1988). In comparison, the solid sedimentary phases preserve higher
Br/I ratios, which overlap planktonic values (Fig. 9.3; Muramatsu et al. 2007).
9.2.4 Iodine Adsorption and Substitution in Carbonate
In addition to I sorption onto organic detritus at the sediment-seawater interface
(Price and Calvert 1973,1977), I is signicantly adsorbed onto the surfaces of a
number of reactive minerals (Ullman and Aller 1985). Available data indicate that
hydrogenetic FeMn crusts formed on igneous substrates, and nodules formed in
sedimentary environments, contain up to 390 ppm F, 1.1 wt% Cl, 200 ppm Br and
900 ppm I which represent a factor of *2 depletion of Cl, a factor of *3
Table 9.2 Representative concentrations of halogens in marine reservoirs
H
2
O
wt%
Cl wt% F ppm Br ppm I ppm
Fluids and sediments
Seawater 96.5 1.93 1.3 67 0.06
Pore uids >95 0.172.3 0.572260 0.1220
Unlithied pelagic sediments
(*including pore uids)
30* 1* 1000 ±300 60 ±60 30 ±25
Lithied marine sediments 700 ±500 1000 ±300 12 ±10 4 ±3
Mineral and organic components
Seaweed <200 ppm 1500
Seaweed (Kelp) 10002000 10,000
Plankton 1000 ±500 10004000 200300
Calcite 300 ±200 <2 ppm
Aragonite 1000 ±500
Mg-carbonate 800 ±500
Fluorapatite 0.62.6
Clay minerals 515 <100 ppm 1000 ±300 (I
adsorbed?)
Mn-crusts/nodules 0.61.1 390 28140 100900
References: Baturin (1988); Carpenter (1969); Fehn et al. (2000,2003,2006); Gieskes and Mahn
(2007); Glasby (1973); Glasby et al. (1978); Hein and Koschinsky (2014); John et al. (2011);
Kennedy and Eldereld (1987a,b); Lu et al. (2008); Mahn and Gieskes (2001); Muinos et al.
(2013); Muramatsu et al. (2001,2007); Muramatsu and Wedepohl (1998); Ohde and Kitano
(1980); Pyle and Mather (2009); Rude and Aller (1991); Tanaka et al. (2013); Tomaru et al.
(2007a,b,c,2009)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 599
Fig. 9.3 Chlorine, Br and I in sediments and sedimentary pore waters. aPore water Br/Cl versus
I/Cl on a log-log scale (with linear scale inset). Pore waters dene an envelope (shaded grey)
extending from seawater (star) and bounded by Br*/I slopes of 0.6 and 1.3 [Br* denotes
seawater-corrected Br where Br* = Br
total
Br
seawater
;Br
seawater
= 0.0035 Cl (Kendrick et al.
2011a)]. In contrast to pore waters, the solid phase in Nankai Trough sediments has a higher Br*/I
of 3 ±1 shown by the yellow envelope.bI/Cl versus salinity expressed as wt% NaCl equivalent.
Pore uid chlorinity is close to seawater in most cases, but is signicantly altered by the formation
and destruction of gas hydrate in some locations. Note that the pore waters have been sampled
from IODP drill holes on both passive and convergent continental margins (see Fehn et al. 2000,
2003,2006; Lu et al. 2008; Mahn and Gieskes 2001; Muramatsu et al. 2001,2007; Tomaru et al.
2007b,c,2009)
600 M.A. Kendrick
enrichment of Br, a factor of *300 enrichment of F and a factor of *2000
enrichment of I compared to seawater (Table 9.2; Glasby 1973; Glasby et al. 1978;
Hein and Koschinsky 2014; Rude and Aller 1991). Hydrogenetic FeMn crusts and
nodules include detrital minerals such as uorapatite and a mixture of MnO
2
and
FeOOH precipitated from cold seawater (with a varying hydrothermal component
in some cases; Baturin and Dubinchuk 2011; Hein and Koschinsky 2014). The
crusts have slow growth rates, very high porosities, and large reactive surface areas
that enable efcient scavenging of reactive trace elements from seawater (Hein and
Koschinsky 2014). The experiments of Ullman and Aller (1985) conrm iodate, but
not iodide, is strongly adsorbed onto Fe-oxyhydroxides and the same may be true
for Mn oxide crusts.
Clay minerals such as montmorillonite [(Na,Ca)
0.33
(Al,Mg)
2
(Si
4
O
10
)
(OH)
2
nH
2
O] have reactive surfaces that adsorb variable quantities of water (e.g.,
nH
2
O) and other substances. Due to the environmental signicance of radioactive
iodide, which is one of the most mobile radioisotopes present in nuclear waste, a
number of investigations have been conducted to evaluate the sorption of iodine to
clay. Experimental results showing signicant iodine sorption have been contro-
versial because of uncertainties regarding the speciation of iodine in the experi-
ments (Glaus et al. 2008), and the potential of iodine to be irreversibly bound in
carbonate formed during the experiment (Claret et al. 2010). However, a recent
study that controlled for these factors reported reversible sorption of iodide onto
clay minerals under controlled PCO
2
(Montavon et al. 2014).
Claret et al. (2010) suggested carbonate within Callovian-Oxfordian clay rock
(France) contains *2 ppm I. Recent experimental data suggest that IO
3
but not I
can substitute for CO
3
2
in calcite (Lu et al. 2010) and the iodate mineral lautarite
[Ca(IO
3
)
2
] is known from I-rich nitrate deposits (Jackson and Ericksen 1997). In
contrast, 2F
can readily substitute for CO
3
2
, but Br
and Cl
(like I
) have low
compatibilities in carbonate (Kitano et al. 1975; Okumura et al. 1986). It should
however, be noted that while the maximum I concentration of 2 ppm reported for
carbonate represents a considerable enrichment over seawater I concentration
of *58 ppb, it is 100 times less than the maximum I concentration of *220 ppm
in sedimentary pore waters (Table 9.2). Further work is clearly required to
understand the interactions of I species, clay minerals and carbonates.
9.2.5 Summary
Two distinct uids that could be advected into the mac portions of the oceanic
crust have been identied and are relatively well dened: (i) seawater and
(ii) sedimentary marine pore water. Sedimentary marine pore waters comprise
seawater modied by the addition of organic components and adsorbed I present in
marine sediments. The majority of pore waters preserve salinities close to seawater,
but salinities of 15 wt% salts can be explained by uid interaction with gas hydrate
(Fig. 9.3). Fluorine has a low solubility in aqueous uids and is retained by or
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 601
incorporated into several marine minerals during diagenesis. In contrast, diagenesis
releases considerable Br and I from the solid phase in organic-rich sediments to the
uid phase. It is usually assumed that Cl has a negligible concentration in the solid
phase of marine sediment (Martin et al. 1993; Muramatsu et al. 2007; Gieskes and
Mahn 2007). However, the few data available for Cl in lithied sedimentary rock
indicate surprisingly high Cl concentrations ranging up to *1000s of ppm, which
might be partly explained by the variable presence of salt derived from seawater or
sedimentary pore waters (Table 9.2; John et al. 2011; Turekian and Wedepohl
1961). Further data are therefore required to evaluate the signicance of water
soluble halogens in marine sediment and to improve the constraint on the Cl content
of the solid phase. Water soluble halogens might be included in estimates of bulk
silicate Earth concentrations, but removal of pore waters during sediment com-
paction suggests that sedimentary pore waters (and therefore salt in dry sediments)
are not relevant for estimation of deep subduction budgets (Peacock 1990).
9.3 The Oceanic Crust-Lithosphere and the Scale of Fluid
Inltration
As a rst step toward evaluating the behaviour of halogens during alteration and
metasomatism of the oceanic crust, it is desirable to review current knowledge
concerning the structure of the oceanic crust and the scale of uid inltration.
It is now recognised that the character of the oceanic crust generated along the
Earths 67,000 km long system of oceanic spreading centres varies as a function of
spreading rate (Bird 2003; Dick et al. 2006; Snow and Edmonds 2007). The classic
Penrosestyle of oceanic crust, as dened by the participants of the 1972 Penrose
conference (Conference Participants 1972), which has a homogenous layered
structure where sediments are underlain by pillow basalts, sheeted dykes, and
layered intrusives, is only generated at fast spreading centres (>60 mm/yr; Dick
et al. 2006). In these cases, the rate of magma supply exceeds the rate of plate
divergence and the seismic Moho at depths of 67 km represents a petrological
transition from layered gabbro to peridotite (Fig. 9.4; Dick et al. 2006).
In contrast, extremely heterogenous crust is generated at ultra-slow spreading
centres, which have rates of divergence of <20 mm/yr, and account for more than
20,000 km of the global ridge system (Bird 2003; Dick et al. 2003). In some cases,
ultra-slow spreading centres are completely amagmatic and characterised by tec-
tonic extension with hydrated ultramac rocks (serpentinites) exposed on the sea-
oor (Dick et al. 2006; Snow and Edmonds 2007). In other cases, a thin and/or
patchy carapace of volcanic rock lies directly on top of layered intrusives or mantle
peridotites (Fig. 9.4; Dick et al. 2006) and the seismic Moho at depths of only
14 km could represent a hydration front between serpentine and mantle peridotite,
as originally suggested by Hess (1962).
602 M.A. Kendrick
Finally, slow spreading ridges, such as the Mid-Atlantic Ridge, with rates of
divergence of between 20 mm/yr and 60 mm/yr, are characterised by alternating
segments of the two crustal styles implying the Penrose model is inadequate for a
large portion of the oceanic crust (Dick et al. 2006; Larsen et al. 2009; Snow and
Edmonds 2007).
9.3.1 Crustal Permeability
Pillow lavas and hyaloclastite breccias in seismic layer 2a of the oceanic crust have
relatively high permeability (*10
13
m
2
), meaning that the upper 500 m of the
crustal basement enjoys good connectivity with cold seawater and is subject to
low-temperature brownstonefacies weathering at very high water/rock ratios
(e.g., w/r > 100; Anderson et al. 1985; Fisher 1998). In situ borehole permeability
Fig. 9.4 Schematic diagram showing two end-member styles of oceanic crust. Penrose style
homogenous oceanic crust with a well dened layered seismic structure is generated at fast
spreading centres (Conference Participants 1972; Dick et al. 2006). Crustal styles vary at slow
spreading centres and may include poorly developed gabbros and dykes or a missing volcanic
carapace; peridotites exposed on the seaoor are hydrated to serpentinites and the seismic Moho
could represent the hydration front. The diagrams are modied after Dick et al. (2006) with
permeability data from Anderson et al. (1985,2012), Fisher (1998), and Fisher and Becker (2000).
Note that fracture permeability continues to variable and poorly dened depths (see text)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 603
measurements indicate permeability decreases to 10
18
10
15
m
2
at greater depths,
but numerical heat ow models imply that localised fracture generated perme-
abilities of up to 10
9
m
2
control the effective permeability in the deeper crust
(Fig. 9.4; Anderson et al. 2012; Fisher 1998; Fisher and Becker 2000; Nehlig
1994). Unlithied sediments represent a low permeability barrier to vertical uid
ow (Fig. 9.4) and hydrothermal systems on the ridge anks are therefore con-
nected to seawater, and each other, through fractures exposed on basement outcrops
(Anderson et al. 2012; Hutnak et al. 2008).
9.3.2 High Temperature Hydrothermal Vent Systems
High temperature (350400 °C) black smoker vent systems, where precipitation of
Fe, Cu, and Zn sulphides generate vent chimneys and sulphide mounds, are found
along mid-ocean ridges, back arc basin spreading centres and at intra-plate sub-
marine volcanoes (Baker and German 2004; Edmond et al. 1979; Hannington et al.
2011; Staudigel et al. 2004; Von Damm et al. 1985). The majority of high tem-
perature hydrothermal elds at spreading centres occur on the ridge axes, but
off-axis systems linked to crustal fractures are known at distances of up to several
km from the axis (Melchert et al. 2008; Rona et al. 1990; Zierenberg et al. 1995).
The circulation of hydrothermal uids is responsible for up to 80% of the
geothermal heat lost in these settings, meaning it is the major mechanism by which
the oceanic crust cools (e.g., Williams et al. 1974).
Heat ow calculations and the distribution of hydrothermal plumes in the water
column indicate that independent vent elds and associated ore deposits (e.g.,
hydrothermal outow zones) occur at intervals of as little as 25 km along fast
spreading centre ridge axes and as far apart as 200 km on slow spreading centre
ridge axes (Baker 2007; Hannington et al. 2011). Seawater is drawn into the
oceanic crust to feed these systems through fractures, gaps between pillow lavas or
dykes, and breccia zones over wide but poorly dened areas between the vent elds
and on the ridge anks (Anderson et al. 2012; Hutnak et al. 2008).
Seismic evidence for brittle deformation suggests that hydrothermal uids
penetrate to maximum depths of *10 km on parts of the Mid-Atlantic Ridge
(Glasby 1998). The TAG hydrothermal eld at 26° N on the Mid-Atlantic Ridge is
situated above a 15 km long detachment fault that penetrates >7 km into the crust
to near Moho depths (de Martin et al. 2007). Periodic reactivation of the fault has
supplied heat to the overlying hydrothermal system and enabled sporadic
hydrothermal activity over tens to hundreds of thousands of years in this area
(Humphris and Cann 2000; Kleinrock and Humphris 1996; Tivey et al. 2003).
In comparison, seismic evidence suggests that hydrothermal uids penetrate
only *3 km into the fast spreading East Pacic Rise (Glasby 1998). Melt lenses at
depths of 13 km on the East Pacic Rise supply heat to the overlying
hydrothermal system but represent a barrier to deeper uid ow. The duration of
hydrothermal activity is difcult to constrain but is commonly related to magmatism
604 M.A. Kendrick
on fast spreading ridges, and individual vent elds are probably sustained over
periods of only tens to hundreds of years (Stakes and Moore 1991).
9.3.3 Serpentinite Hosted Systems
The Lost City Hydrothermal Field on the Atlantis Facture Zone occurs 15 km from
the Mid-Atlantic ridge axis at 30° N (Kelley et al. 2001). This system is hosted by
ultramac rocks and is characterised by carbonate vent chimneys and low temper-
ature (4090 °C) alkaline uids (pH 910) (Kelley et al. 2001,2007). The alkalinity
of the uids and the presence of hydrogen, methane and abiogenic hydrocarbons are
characteristics of uids produced by serpentinisation of ultramac lithologies
(Kelley et al. 2007; Proskurowski et al. 2008). The discovery of this system was
signicant because it is driven by heat generated by exothermic serpentinisation
reactions, rather than magmatism, and similar reactions could potentially drive
numerous hydrothermal systems distal to ridge axes, implying a much higher level
of seaoor hydrothermal activity than previously recognised (Kelley et al. 2001).
Serpentinites are commonly exposed along transform faults and low angle detach-
ments related to oceanic core complexes, providing evidence for widespread uid
activity at varying distance from oceanic spreading axes (Fryer 2002).
The slab bend formed immediately before subduction is now recognised as an
additional site of important hydrothermal alteration (Fig. 9.5). Seismic evidence
suggests faults cut at least 20 km below the seaoor in this location, with the
Fig. 9.5 Schematic diagram highlighting the slab bend as an important site for deep (>20 km)
fractures and hydrothermal alteration of the lithospheric mantle (see text). Note that fractures at the
slab bend provide a possible pathway for sedimentary pore waters to inltrate the lithosphere
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 605
potential for large scale serpentinisation of the lithospheric mantle (Fig. 9.5; Ranero
et al. 2003). Heat ow measurements off shore Nicaragua and Central Chile indi-
cate these areas are characterised by unexpectedly low heat ow consistent with
heat loss by hydrothermal activity (Grevemeyer et al. 2005).
9.4 Black Smoker Vent Fluid Chemistry and Halogens
In comparison to seawater, high temperature black smoker vent uids contain
negligible Mg or SO
4
2
(dissolved sulphur is S
2
), and are often depleted in U, P
and F (Fig. 9.6). In contrast, they are enriched in trace elements, dissolved SiO
2
and
contain 510 times more K and Ca relative to Na, than seawater (Fig. 9.7). Typical
vent uids have pH of 34, maximum temperatures of close to 400 °C and salinities
that range from 0.1 to 6 wt% salts (Fig. 9.8).
The composition of hydrothermal vent uids results from chemical reaction of
seawater with the oceanic crust and the mobilisation of trace elements and silica made
possible by the high temperature of the uid, acidity, and presence of the chloride
ligand (Bischoff and Dickson 1975; Humphris and Thompson 1978; Seyfried and
Bischoff 1981). The take up of Mg into Mg-rich hydrous-silicate minerals, including
smectite, chlorite, amphibole, and talc, is important because Mg-metasomatism is the
primary source of vent uid acidity and leads to extensive hydration of the crust
(reaction 9.2; Bischoff and Dickson 1975; Seyfried and Mottl 1982).
Fig. 9.6 Comparison of selected elements in seawater and Hanging Garden vent uids from 21°
N on the East Pacic Rise (Von Damm et al. 1985; Li and Schoonmaker 2003). The diagram
shows that relative to the 1:1 line (dashed), most trace elements are strongly enriched in the vent
uids, but that S, F, and P are slightly depleted and Mg is quantitatively removed from the vent
uids. Note that the Hanging Garden Vent uids have near seawater salinity and that this diagram
uses a log scale meaning that small variations in concentration (e.g., a factor of 2) cannot be easily
evaluated (see Fig. 9.7)
606 M.A. Kendrick
3MgCl2ðaqÞþ4SiO2ðaqÞþ4H2OðlÞ¼Mg3Si4O10 OHðÞ
2ðsÞþ6HClðaqÞð9:2Þ
SO
4
2
is removed from solution by reduction to sulphide and precipitation with
Ca as anhydrite; Ca is released to the uid by albitisation of plagioclase and the Ca/
Na ratio of the uid is subsequently controlled by plagioclase and epidote solid
solution (Berndt et al. 1989). Potassium is incorporated into K-minerals at low
temperature but it is released to the uid at temperatures of >150 °C, which means
most high temperature vent uids have K/Cl ratios higher than seawater. The K/Cl
ratio of the vent uid is usually limited by the low K content of the basaltic crust,
but relatively high K/Cl ratios of *0.3 occur in the Manus basin vent uids where
more evolved K-rich rocks are present on the seaoor (Fig. 9.7; Reeves et al. 2011).
The growth of Mg-rich hydrous minerals causes a nite increase in the salinity
of hydrothermal uids; however, this is only detectable at low water-rock ratios of
<1 (e.g., Kelley and Robinson 1990; Seyfried et al. 1986; Seyfried and Mottl 1982).
Therefore, the occurrence of vent uids with salinities of less than seawater, and in
some cases the dominance of vent uids with lower than seawater salinity over
many years of observation (Campbell et al. 1988), has been interpreted as strong
evidence for phase separation and segregation of brines and vapours in submarine
hydrothermal systems (Bischoff and Rosenbauer 1989; Buttereld et al. 1997;
Seyfried et al. 2003; Von Damm et al. 1997) (Table 9.3).
Fig. 9.7 Ca/Na versus K/Na weight ratios of vent uids for which Cl and Br data are available
(below). Vent uids are variably enriched in K and Ca, relative to Na in seawater; the seawater
ratios are shown by the star (Buttereld et al. 1990; Ishibashi et al. 1994; James et al. 1995,2014;
Mottl et al. 2011; Pester et al. 2012; Reeves et al. 2011; Seyfried et al. 2003; Von Damm 1997,
1998,2000,2003)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 607
9.4.1 Phase Relations in the H
2
O-NaCl System
The phase relations of sub-marine hydrothermal systems are strongly controlled by
the presence of salts. Pure water exists as a homogenous supercritical uid above its
critical point (374 °C; 220 bars). In contrast, a critical curve extends beyond the
critical point of pure H
2
O in any binary H
2
O-salt system, and the two-phase region
is vastly expanded (Fig. 9.9; e.g., Bischoff and Pitzer 1989; Driesner and Heinrich
2007). The critical point of seawater (407 °C; 220 bars) lies on the critical curve of
Fig. 9.8 Temperature and salinity data for selected vent uids. Note that the data show maximum
vent temperature and that most of the salinities are for end-member hydrothermal uids with zero
Mg (e.g., uncontaminated by seawater). A conjugate brine and vapour pair emitted from vent F at
9° N on the East Pacic Rise is indicated (below; Von Damm et al. 1997 the other data are from
Buttereld et al. 1990; Edmond et al. 1979; Ishibashi et al. 1994; James et al. 1995,2014;
Kawagucci et al. 2011; Mottl et al. 2011; Pester et al. 2012; Reeves et al. 2011; Seyfried et al.
2003, 2011; Von Damm et al. 1997,1998,2000,2003,2005; Wu et al. 2012)
Table 9.3 Halogen signatures of uids in the oceanic lithosphere
Salinity F/Cl Br/Cl I/Cl
Seawater 3.5 0.00007 0.0035 0.000003
Sedimentary pore uids 15 0.000050.0002 SW
a
0.014 SW
a
0.04
Vent uids 0.17 0.0000060.0005 0.0010.005 SW
a
0.001
Crustal brines (uid inclusions) 1060 <0.0001 SW
a
0.005 SW
a
0.00012
SW
a
= seawater
References are given in Table 9.1 and Figs. 9.3,9.11 and 9.15
608 M.A. Kendrick
the H
2
O-NaCl system and joins the boiling curve to the condensation curve in
Fig. 9.9a (e.g., Bischoff and Pitzer 1989; Bischoff and Rosenbauer 1984; Driesner
and Heinrich 2007). Sub-critical phase separation occurring below the critical point
of seawater produces a small quantity of vapour, whereas super-critical phase
separation occurring above the critical point of seawater produces a small quantity
of brine (Fig. 9.9).
The slope of the critical curve implies that phase separation will produce pro-
gressively more saline brines at higher pressures and temperatures deeper within the
oceanic crust (Fig. 9.9). The salinity of brines also increases as the L-V isotherms
are overstepped, such that at 500 °C and 540 bars the rst brine exsolved has a
salinity of 28 wt%, but further decompression to 450 bars would yield a brine of 44
wt% NaCl (points 1 and 2; Fig. 9.9b). It is important to note that the salinities of the
vapours and brines produced at different temperature and pressure conditions are
controlled exclusively by the shape of the liquid-vapour envelope. For example at
500 °C and 450 bars, parental uids with salinities of either 3.2 wt% NaCl
(seawater) or 10 wt% NaCl, would both exsolve into conjugate uids with identical
Fig. 9.9 Phase diagrams for the binary H
2
O-NaCl system. aPressure versus temperature and
bpressure versus wt% NaCl. Note that the approximate depth of different crustal units under
hydrostatic pressure is shown and that the wt% NaCl scale switches from log to linear at 10 wt%.
Part a shows the critical curve, with the critical point of seawater estimated from 3.2 wt% NaCl
solution, and the L-V-S boundary that separates the vapour-salt and vapour-liquid elds. Part
bshows the critical curve, liquid-salt eld, and isotherms that dene the two phase eld at selected
temperatures. The effects of depressurisation are shown for 4 different pressure-temperature
regimes (see text). Data and equations used to construct the gures are from Bischoff and Pitzer
(1989) and Driesner and Heinrich (2007)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 609
salinities, but at different proportions that can be estimated using the lever rule
(point 2; Fig. 9.9b).
In addition to the well documented phenomena of liquid-vapour immiscibility,
precipitation of hydrothermal-halite is possible under some circumstances (e.g.,
point 4; Fig. 9.9). This is possible for seawater, which is approximated by the
H
2
O-NaCl system (Fig. 9.9), because hydrostatic pressures above spreading centres
at typical water depths of 20002500 m are only *200250 bars and seawater
heated above *400 °C would separate into steam and halite (Fig. 9.9). Buttereld
et al. (1997) presented photomicrographic evidence for halite coating glass surfaces
and intergrown with TiO
2
, AlO(OH), and sphalerite, that was interpreted as evi-
dence for precipitation of hydrothermal halite during high temperature interaction
of seawater with seaoor lavas. The precipitation of hydrothermal halite has also
been suggested as a possible cause of Br and Cl fractionation in some hydrothermal
vent uids (Berndt and Seyfried 1997; Foustoukos and Seyfried 2007). However, it
should be noted that the vapour-salt eld is probably restricted to much shallower
depths in the multi-component salt systems that are relevant for hydrothermal uids
than in the H
2
O-NaCl system that is relevant for seawater (below).
9.4.2 Complex Salt Systems and Hydrolysis
Hydrothermal vent uids are better modelled in the ternary H
2
O-NaCl-CaCl
2
system than in the NaCl-H
2
O system (Vanko et al. 1988) because they have Ca/Na
and K/Na ratios elevated by 510 times compared to seawater (Fig. 9.7). The two
phase eld is even larger in these complex salt systems than in the H
2
O-NaCl
system. The critical curve for H
2
O-K/Cl is similar to that of H
2
O-NaCl, but the
critical curve for H
2
O-CaCl
2
moves to signicantly higher pressure at any given
temperature than in the H
2
O-NaCl system (Fig. 9.10; Bischoff et al. 1996; Driesner
and Heinrich 2007; Hovey et al. 1990). In addition, at temperatures of 400450 °C,
the vapour-salt eld shrinks back from pressures of 180250 bars in the H
2
O-NaCl
system (that overlap the seaoor; Fig. 9.10), to 130170 bars in the H
2
O-KCl
system and to as little as *4065 bars in the H
2
O-CaCl
2
system (Fig. 9.10;
Driesner and Heinrich 2007; Hovey et al. 1990; Keevil 1942; Ketsko et al. 1984).
In addition to the role of Cl as an important ligand for metal transport in
hydrothermal solutions (e.g., Yardley 2005), hydrolysis reactions in H
2
O-salt
systems can have a signicant inuence on the pH of hydrothermal uids (e.g.,
reaction 9.3; Bischoff et al. 1996; Seyfried et al. 1988; Seyfried and Bischoff 1981).
Hydrolysis reactions in binary salt systems have the general form:
xH2OþMxþCl
x¼xHCl þMxþOHðÞ
xð9:3Þ
The equilibrium lies to the left under Earths surface conditions, meaning salt
solutions have a neutral pH, but the equilibrium can be signicantly shifted to the
right under some crustal conditions. Bischoff et al. (1996) demonstrated that at
610 M.A. Kendrick
temperatures of 350500 °C hydrolysis is signicantly more important in the
CaCl
2
-H
2
O system than in the binary NaCl-H
2
O system. Fournier and Thompson
(1993) reported that hydrolysis does not occur in the liquid-vapour region of the
H
2
O-NaCl-KCl system, but that HCl is a signicant component of the steam in
equilibrium with halite in the vapour-salt region of the H
2
O-NaCl and
H
2
O-NaCl-KCl systems. The varying degrees of hydrolysis in the different salt
systems are related to the relative solubilities of the metal hydroxide produced by
hydrolysis: Ca(OH)
2
has a low solubility meaning it is removed from solution more
efciently than NaOH and as a result reaction 9.3 is pushed further to the right
(Bischoff et al. 1996).
In nature, the hydroxide component produced by hydrolysis reacts with silicates
to form hydrous silicate minerals in the oceanic crust. For example, hydrolysis of
MgCl
2
leads to Mg being xed in hydrous silicate minerals under prograde con-
ditions as seawater is drawn into hydrothermal systems (reaction 9.2; Seyfried et al.
1988). The reactivity of Mg in the oceanic crust means that it is quickly removed
from solution and Mg is not present in the hydrothermal uids that reach the high
temperature reaction zone above the magma chamber. In the absence of MgCl
2
,
hydrolysis of CaCl
2
becomes the main acid generating reaction and is important in
the hydrothermal upow zones; hydrolysis of NaCl can be locally important in
zones of Na metasomatism (Seyfried et al. 1988). The generation of acidity by
hydrolysis reactions is important because it enables mobilisation of heavy metals
Fig. 9.10 Phase diagrams for various binary salt systems. aPressure versus temperature plot
showing the critical curves for H
2
O-NaCl and H
2
O-CaCl
2,
and the L-V-S curves for the H
2
O-
NaCl, K/Cl, and CaCl
2
systems. Note that the vapour-salt eld shrinks back from a maximum size
in the H
2
O-NaCl system to a minimum size in the H
2
O-CaCl
2
system. bPressure versus molar
equivalent CaCl
2
, HCl, and Ca(OH)
2
. The binary H
2
O-CaCl
2
system differs to the H
2
O-NaCl
system in that hydrolysis occurs under high temperaturelow pressure conditions (see text). Data
and equations used to construct the gures are from Bischoff and Pitzer (1989), Bischoff et al.
(1996), Driesner and Heinrich (2007), Hovey et al. (1990) and Ketsko et al. (1984)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 611
from seaoor lithologies (Seyfried and Bischoff 1981). Furthermore, HCl partitions
into the vapour produced by phase separation (Fig. 9.10; Bischoff et al. 1996),
meaning segregation of vapours and brines enables discrete zones of alkali or acid
metasomatism to develop in different parts of the oceanic crust, or waves of acid
metasomatism to pass through a hydrothermal system (Kigai and Tagirov 2010).
9.4.3 Halogen Abundance Ratios in Vent Fluids
There has been considerable interest in the relative abundances of halogens in
hydrothermal vent uids. The existing data show vent uids have F concentrations
ranging from *0.1 to 7 ppm, compared to the seawater value of 1.3 ppm F, and
that F varies independently of Cl with vent uid F/Cl ratios varying by almost two
orders of magnitude around the seawater value (Fig. 9.11a; Edmond et al. 1979;
Fig. 9.11 The halogen systematics of hydrothermal vent uids: aF/Cl versus Br/Cl; bBr/Cl
versus salinity; and cI/Cl versus Br/Cl (Buttereld et al. 1990; Campbell and Edmond 1989;
Campbell et al. 1994; Edmond et al. 1979; Gieskes et al. 2000;2002; Ishibashi et al. 2002; James
et al. 1999,2014; Kawagucci et al. 2011; Mottl et al. 2011; Oosting and Von Damm 1996; Reeves
et al. 2011; Seyfried et al. 2003; Von Damm et al. 1985,1997,1998,2000,2003,2005; Wu et al.
2012; You et al. 1994)
612 M.A. Kendrick
Gieskes et al. 2000,2002; James et al. 2014; Mottl et al. 2011; Reeves et al. 2011;
Von Damm et al. 1985). The reported variation in F ts well with experimental
studies that demonstrate F is removed from solution as seawater interacts with
basalt and is heated from 150 to 250 °C, but re-enters solution with further heating
to 300 °C (Seyfried and Ding 1995). Fluorine is incorporated into Mg-carbonate
(cf., Fig. 9.1) and hydrous minerals at low temperature but the direction of
hydroxyl-uoride and carbonate-uoride exchange reactions reverses at high tem-
perature when the speciation of dissolved F changes from F
(aq)
to HF
(aq)
, which
reduces the chemical activity of the F
anion (Seyfried and Ding 1995).
Consequently the experimental studies imply that while magmatic F might be
mobilised in deeper parts of the crust as HF
(aq)
, it is likely that seawater F
(aq)
is
removed from solution during heating and is not advected deeply into the crust
(Seyfried and Ding 1995). The co-depletion of some hydrothermal uids in P as
well as F suggests that precipitation of uorapatite is also important (Fig. 9.6;
Gieskes et al. 2002).
A number of studies have investigated Br in hydrothermal uids and demon-
strated that the Br/Cl ratio of the vent uids is usually within 10% of the seawater
value and that it does not vary systematically as a function of salinity (Fig. 9.11b).
Furthermore, conjugate vapours and brines from vent F at 9°16N on the East
Pacic Rise (Von Damm et al. 1997) and Milos Island in the Hellenic Arc (Wu
et al. 2012) have similar Br/Cl ratios (Fig. 9.11b). Taken together these data
strongly suggest that the Br/Cl ratio is not strongly or systematically altered by
liquid-vapour phase separation in the majority of hydrothermal systems investigated
(Fig. 9.11b). This conclusion is consistent with experiments on NaCl solutions that
show no systematic fractionation of Br and Cl during sub-critical phase separation
(Berndt and Seyfried 1997). Experiments under other conditions have shown
fractionation of Br relative to Cl into either vapours or brines during phase sepa-
ration (see Berndt and Seyfried 1990; Foustoukos and Seyfried 2007; Liebscher
et al. 2006). However, the results from the different laboratories are not in agree-
ment and the available Br/Cl data summarised for vent uids in Fig. 9.11 are more
easily explained by the conservation of Br/Cl during phase separation with unusual
Br/Cl ratios of uids in specic settings being explained by uid-rock reactions
(below).
The unusually high Br/Cl ratios in uids from a small number of vents
(Fig. 9.11b) can be explained by uid interaction with sedimentary material,
because where Br/Cl and I/Cl data are both available these ratios co-vary in a
manner comparable to that observed for sedimentary pore waters (Fig. 9.11c;
Campbell and Edmond 1989; Campbell et al. 1994; Gieskes et al. 2002; Kawagucci
et al. 2011; Mottl et al. 2011; You et al. 1994). Furthermore, I concentrations are
correlated with NH
4
and other sediment-derived components in these uids
(Campbell et al. 1994; You et al. 1994). Note that limited data are available for I or
F in vent uids and halogen data for sedimentary pore uids with mantle-like
3
He/
4
He signatures, which are believed to represent hydrothermal uids from
beneath the Escanaba Trough hydrothermal system (Hole 1038B-H; Ishibashi et al.
2002; James et al. 1999), are included in Fig. 9.11 as light blue symbols (Gieskes
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 613
et al. 2000,2002). These hydrothermal pore uidshave Br/Cl and I/Cl values
overlapping the range reported for Escanaba Trough vent uids (Fig. 9.11c;
Campbell and Edmond 1989; Campbell et al. 1994; You et al. 1994).
Unusually low Br/Cl ratios are restricted to a few locations between 9° and 10°
N on the East Pacic Rise (Oosting and Von Damm 1996) and Milos Island in the
Hellenic Arc (Fig. 9.11b; Wu et al. 2012). It is possible that vent uids acquired
low Br/Cl ratios in these settings by uid interaction with sedimentary halite or by
leaching glassy magmatic rocks that are typically characterised by Br/Cl of less than
seawater (see mantle eld in Fig. 9.11c; Jambon et al. 1995; Kendrick et al. 2013a,
2017; Schilling et al. 1980). Similarly low Br/Cl ratios have been reported for
unusual sediment pore waters that have interacted with glassy volcaniclastic rocks
in the Aoba Basin of the New Hebrides convergent margin (Martin 1999). Strongly
altered, palagonitised glasses can also be depleted in Cl and Br relative to pristine
glasses (Fig. 9.12; Kendrick et al. 2015a) conrming that these elements are
mobilised during low temperature alteration of glass.
Finally, the fact that most vent uids have Br/Cl ratios within 10% of seawater
(Fig. 9.11) implies that in most cases, uid-rock ratios are sufciently high that Br
and Cl behave conservatively in hydrothermal vent uids and are not signicantly
altered by phase separation or reaction with mac lithologies (Br concentrations are
more readily altered by interaction with sediments; Fig. 9.3; Muramatsu et al. 2007;
Price and Calvert 1977). However, leaching of Cl > Br from glassy rocks is pos-
sible and has the potential to lower the Br/Cl ratio of a uid. Furthermore, at low
water-rock ratios, precipitation of OH > Cl > Br in Fe/Mg-hydroxy chlorides, or
amphibole, has the potential of increasing a uids salinity and Br/Cl ratio
(Kendrick et al. 2015b; Seyfried et al. 1986; Svensen et al. 1999,2001; Vanko
1986). Therefore, it should not be assumed that Br and Cl behave conservatively
under all hydrothermal conditions.
Fig. 9.12 Halogens in
pristine glasses and their
weakly altered or
palagonitised counterparts
from Macquarie Island and
Samoa (Kendrick et al. 2012a,
2015a)
614 M.A. Kendrick
9.5 Fluids in the Deeper Crust
Evidence for the nature of uids in deeper parts of the oceanic crust comes from
mineral alteration (Gillis and Meyer 2001; Vanko 1986), uid inclusions (Fig. 9.13;
Kelley et al. 1992,1993; Kendrick et al. 2015b; Nehlig 1991; Vanko 1988; Vanko
et al. 1992), and studies of magmatic glass that contain assimilated halogens (below;
Kendrick et al. 2013a,2015a,2017). Fluid inclusions have been investigated in
minerals such as quartz, anhydrite, sphalerite, epidote, amphibole, and chlorite from
a number of dredges and ophiolites. These include samples from within and just
below the vent chimney all the way down to upper amphibolite facies metagabbros
from layer 3 of the oceanic crust (Castelain et al. 2014; Juteau et al. 2000; Kelley and
Delaney 1987; Kelley and Robinson 1990; Kelley et al. 1992,1993;Lécuyer et al.
1999; Nehlig 1991; Nehlig and Juteau 1988; Vanko 1986,1988,1992,2004).
Fig. 9.13 Photomicrographs
of vapour and brine uid
inclusions in a quartz vein
hosted by amphibolite
metagabbro from layer 3 of
the crust. This sample (745),
dredged from the
Mathematician Ridge of the
Pacic Ocean, has been
described in detail by Vanko
(1988) and Kendrick et al.
(2015b). The vapour bubbles
have a high degree of ll in
the vapour inclusions. In
contrast, brine inclusions have
much smaller vapour bubbles
and visible halite daughter
minerals which have a bright
green hue and cubic form
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 615
Collectively these studies demonstrate the predominance of low salinity
Na-Ca-Cl uids throughout the oceanic crust that are broadly similar to vent uids.
For example, Nehlig (1991) reported that more than 95% of the uid inclusions
examined in samples taken from crustal sections through the Semail and Trinity
ophiolites of Oman and California, and in samples from the Gorringe Bank of the
East Pacic Rise, had salinities within 1 wt% of seawater. However, in addition to
the dominant uid inclusions with salinities of 0.16 wt%, which are similar to vent
uids, uid inclusions with extreme Ca enrichment (Ca/Na = 1), which is stronger
than known from vent uids (Fig. 9.7), have been reported (Vanko et al. 1992) and
brine inclusions with salinities of up to *50 wt% salts appear to be a common minor
component at all levels of the oceanic crust (Fig. 9.13; Juteau et al. 2000; Kelley and
Delaney 1987;Lécuyer et al. 1999; Vanko 1988,2004; Aranovich et al. 2015).
Brine inclusions with salinities of *30 wt% salts have been reported in the vent
chimneys of several systems in the Lau Basin and within a few hundred metres of
the seaoor beneath the Pacmanus system of the Manus Basin, demonstrating that
high salinity brines are not restricted to deep parts of the oceanic crust (Lécuyer
et al. 1999; Vanko et al. 2004). However, high salinity brine inclusions are most
commonly reported from deeper amphibolite settings (Vanko 1988; Kendrick et al.
2015b) and above newly emplaced gabbroic intrusions (Kelley and Delaney 1987)
or plagiogranites that form in the roof zones of crustal magma chambers (Kelley
and Robinson 1990; Nehlig 1991; Vanko et al. 1992). In many cases the brine
inclusions co-exist with more abundant low salinity vapour inclusions that provide
evidence for their origin by phase separation (e.g., Juteau et al. 2000; Kelley and
Delaney 1987; Vanko 1988). The dominance of vapour inclusions is the expected
result of seawater undergoing phase separation (e.g., point 2 in Fig. 9.9; Bischoff
and Pitzer 1989). However, the vapour inclusions can have salinities that are even
greater than seawater suggesting complex multi-stage histories and/or input of a
magmatic component.
In some cases, brine inclusions dominate samples recovered from layer 3 of the
crust. These inclusions have been interpreted as brines that have been segregated
from their conjugate vapours, but their origin by phase separation cannot be proven
(Kelley and Robinson 1990; Vanko et al. 1992). The uid inclusion evidence for
possible segregation of brines in layer 3 of the oceanic crust is signicant because
brine segregation and double diffusive convection of uids in the oceanic crust has
previously been invoked to explain the predominance of uids with lower than
seawater salinity emitted from vents on the seaoor (Fig. 9.14; Bischoff and
Rosenbauer 1989). It is suggested the brines would be segregated because of the
different wetting properties and buoyancies of vapours and brines in a micro-scale
fracture network. Buoyant vapours would be preferentially lost to the overlying
hydrothermal system, whereas dense brines could be trapped in the lower crust. If
correct, this model implies brines generated by multiple episodes of phase sepa-
ration gradually accumulate in the deeper crust and they are implied to have long
residence times in the crust at very low effective water-rock ratios (Fig. 9.14). The
ultimate fate of brines within this model is unknown but brines might eventually be
ushed out of the crust by advection of more typical low salinity uids in the
616 M.A. Kendrick
waning stages of hydrothermalism or they might be transported laterally and leak
out of the ridge anks away from the vents (Bischoff and Rosenbauer 1989). In
addition, a portion of these brines could be assimilated by magmas driving the
hydrothermal system (Kendrick et al. 2013a,2015a,2017) or involved in ux
melting and the genesis of plagiogranites (Aranovich et al. 2010,2015).
A major uncertainty related to the brine uid inclusions is whether they originate
purely from evolved seawater or include a component derived from late-stage
magmatic uids. Magmatic uids exsolved from basic rocks are dominated by CO
2
with a low salinity aqueous component (Dixon et al. 1995; Webster et al. 1999).
However, more saline uids might sometimes be exsolved from evolved intrusive
rocks such as plagiogranites in magma chamber roof zones (Kelley and Robinson
1990; Kelley et al. 1992). If magmatic uids enter the two-phase eld, their salinity
depends on the shape of the liquid-vapour envelope in the relevant water-salt
system (e.g., Fig. 9.9 and 9.10), meaning that it is not possible to infer the source of
the uid (e.g., seawater or magmatic) from the measured salinity. Vanko et al.
(1992) suggested that a minor magmatic component could be present in high
salinity CO
2
-bearing uid inclusion assemblages from the Oceanographer
Transform, but favoured an origin entirely from evolved seawater for brines trapped
without accompanying CO
2
on the Mathematician Ridge (Vanko 1988,1992;
Kendrick et al. 2015b). Reeves et al. (2011) suggested that unusually high F/Cl
Fig. 9.14 Schematic diagram illustrating the double diffusive convection concept of Bischoff and
Rosenbauer (1989). This model predicts brines are segregated from vapours and preferentially
retained in deeper portions of the crust. The presence of brines may help explain some styles of
alteration and Cl-rich amphiboles. The ultimate fate of the brines may be to be diluted by lower
salinity uids and ushed out of the crust, to migrate laterally and leak from ridge anks or be
assimilated by magmas (Fig. 9.16)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 617
ratios in Manus Basin vent uids could result from the input of magmatic volatiles
(Fig. 9.11).
9.5.1 Halogens in Brines and Assimilation by Magmas
The systematics of Br/Cl, I/Cl, and F/Cl in brine inclusions can be inferred from
two separate lines of inquiry. Kendrick et al. (2015b) investigated halogens and
noble gases in high salinity brine and vapour uid inclusions in six quartz/epidote
veins from the Mathematician Ridge of the NE Pacic using a bulk extraction
technique. The uid inclusions in all the veins were shown to have a similar range
of Br/Cl and I/Cl irrespective of the relative proportions of brine and vapour uid
inclusions (Fig. 9.15). In addition, Br/Cl was not correlated with Ar/Cl, which is
strongly fractionated by phase separation (Kendrick et al. 2015b). Taken together
these data suggest that Br/Cl and I/Cl were not fractionated during phase separation
of Mathematician Ridge uids, consistent with the behaviour of Br/Cl in vent uids
(Sect. 9.4.3). On average the Mathematician Ridge uids have Br/Cl *20% higher
than seawater and I/Cl intermediate of seawater and the mantle (Fig. 9.15; Kendrick
et al. 2015b). These compositions were attributed to mixing halogens introduced by
seawater with halogens mobilised from the crust and preferential exclusion of Br
and I relative to Cl from amphibole crystallised at low water/rock ratios (Fig. 9.15;
Kendrick et al. 2015b). Proton Induced X-ray Emission (PIXE) has been used to
analyse some individual brine inclusions from the Mathematician Ridge,
Oceanographer Transform and Oman ophiolite (Juteau et al. 2000; Vanko et al.
2001). The majority of the uid inclusions investigated by PIXE are indicated to
Fig. 9.15 Halogens in uid
inclusions and amphibole
separated from Mathematician
Ridge amphibolites and
metagabbros. Fluid inclusions
where analysed by crushing
vein minerals (quartz and
epidote) and the amphibole
wall-rock in vacuum
(Kendrick et al. 2015b). Note
that the range of uid
inclusion compositions is very
similar to the composition of
brines assimilated by
submarine magmas dredged
from the Lau Basin,
Galapagos Spreading Centre
and Samoa (Fig. 9.16;
Kendrick et al. 2013a,2015a)
618 M.A. Kendrick
have Br/Cl close to or above seawater, but the precision of the analyses is low
(Juteau et al. 2000; Vanko et al. 2001).
The second constraint on the halogen composition of high salinity brines in the
oceanic crust comes from recent studies of volcanic glass (Kendrick et al. 2013a,
2015a,2017). It has long been recognised that some volcanic glasses contain
excess Clintroduced by the assimilation of seawater-derived components
(Michael and Cornell 1998; Michael and Schilling 1989). Most workers have
favoured the assimilation of hydrothermally altered crust to account for excess Cl.
However, plotting the H
2
O/Cl and K/Cl ratios of all known glasses, identied as
containing excess seawater-derived Cl (Coombs et al. 2004; Freund et al. 2013;
Kent et al. 1999a,b,2002; le Roux et al. 2006; Lytle et al. 2012; Wanless et al.
2010), shows that Cl is introduced into these melts by high salinity brines (e.g.,
Fig. 9.16; Kendrick et al. 2013a). This is shown by the data in Fig. 9.16 which
Fig. 9.16 Halogens in magmatic glasses affected by brine assimilation. aK/Cl versus H
2
O/Cl;
bF/Cl versus H
2
O/Cl; cBr/Cl versus K/Cl; and dBr/Cl versus I/Cl (Kendrick et al. 2013a,2015a;
le Roux et al. 2006; Lytle et al. 2012). The mantle eld is dened by Kendrick et al. (2013a,2014,
2015a,2017). Brines have K/Cl and F/Cl of <0.1 (e.g., Fig. 9.8) and H
2
O/Cl that is proportional to
salinity (parts aand b). The altered ocean crust (AOC) is assumed to have low K/Cl and high H
2
O/
Cl based on existing data and the relatively low compatibility of Cl in hydrous minerals compared
to H
2
O (Ito et al. 1983; Sano et al. 2008)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 619
extend from a range of compositions representative of the mantle to converge on a
single component with very low ratios of H
2
O/Cl, K/Cl, and F/Cl that are char-
acteristic of a Na-Ca-K-Cl brine with salinity of >50 wt% salts (Fig. 9.16; Kendrick
et al. 2013a,2015a,2017). In contrast, altered crustal material would have much
higher H
2
O/Cl (Ito et al. 1983). The Br and I data available for glasses, affected by
brine assimilation from the NW part of the Lau Basin, the Galapagos, and Samoa,
dene coherent mixing trends with K, Cl, and H
2
O, demonstrating that the
assimilated brines are characterised by Br/Cl and I/Cl ratios a few 10s of percent
higher than seawater (Fig. 9.16), that are very similar to the range of compositions
determined for Mathematician Ridge uid inclusions (Fig. 9.15; Kendrick et al.
2015b).
The brine assimilation data are signicant because: (i) some of the glasses
affected have high concentrations of H
2
O and CO
2
, which demonstrate that
assimilation must have occurred at depth in crustal magma chambers rather than on
the seaoor (Coombs et al. 2004; Kendrick et al. 2013a; le Roux et al. 2006), and
(ii) the data provide evidence that seawater-derived brines, with Br/Cl and I/Cl
distinct from the mantle, not only penetrate the lower crust but come into direct
contact with, and are assimilated by crustal magmas (Fig. 9.16). Mass balance
calculations suggest the affected magmas from the NW part of the Lau Basin, the
Galapagos, and Samoa assimilated up to 0.5% of their total mass in brine which
introduced 070% of their total Cl and 030% of their total H
2
O (Kendrick et al.
2013a,2015a,b). Therefore, assimilation of brines generated in hydrothermal
systems can alter the concentrations of halogens and relative abundance ratios (F/
Cl, Br/Cl and I/Cl) of newly forming crust (e.g., magma), even before the new crust
has been accreted. In addition, the ductile zone, surrounding the magma chambers,
is not a complete barrier to assimilation of seawater-derived volatiles.
9.6 Halogens in Altered Oceanic Crust and Lithosphere
A relatively small number of studies have investigated the bulk halogen content of the
altered ocean crust with the majority of studies focused on Cl (Barnes and Cisneros
2012; Bonifacie et al. 2007; Chavrit et al. 2016; Floyd and Fuge 1982; Ito and
Anderson 1983; Ito et al. 1983; Kendrick et al. 2015b; Magenheim et al. 1995; Sano
et al. 2008). The earliest estimates for whole rock Cl concentration were made by
electron microprobe measurements of Cl in individual mineral phases, followed by
estimation of the modal abundances of the minerals in the rock (Ito et al. 1983). Major
problems with this approach are that intra-granular Cl and uid inclusion hosted Cl,
which can account for 100200 ppm Cl in the bulk sample (Kendrick et al. 2015b), or
Cl present in volumetrically minor phases such as Fe hydroxychlorides are not
included in the analysis. In addition, although detection limits of better than 50 ppm
Cl can be achieved with sufciently long counting times (e.g., Michael and Cornell
1998) most minerals have Cl concentrations of less than the typical 100300 ppm
detection limits for Cl achieved by routine electron microprobe analysis.
620 M.A. Kendrick
Direct measurements of Cl have been achieved by various instrumental neutron
activation techniques that provide modest 2rprecision of 1520% (Sano et al.
2008) or 200 ppm Cl (Barnes and Cisneros 2012), but have the signicant
advantage of requiring minimal sample processing.
Direct measurements of F, Cl, and I have also been achieved by digestion of
powders and spectrophotometry (Floyd and Fuge 1982) and more recently by
combining pyrohydrolysis for halogen extraction from powders with ion chro-
matography for F and Cl analysis (Bonifacie et al. 2007; John et al. 2011;
Magenheim et al. 1995; Sharp and Barnes 2004) and ICP-MS for Br and I analysis
(John et al. 2011). These techniques can provide accurate results with internal
precision of better than 5%. However, Bonifacie et al. (2007) demonstrated that Cl
yields during pyrohydrolysis can vary between labs with yields as low as 40% in
some of the early studies, implying that signicant unquantied fractionation of
halogens during extraction is possible.
Kendrick et al. (2011a,b,2013a,2015b) and Chavrit et al. (2016) employed the
noble gas methodfor analysis of Cl, Br, I, and K in rock chips and mineral separates
whereby neutron irradiation is used to generate noble gas proxy isotopes for the
halogens, which can be precisely measured by noble gas mass spectrometry (Böhlke
and Irwin 1992; Johnson et al. 2000; Kendrick 2012). This technique combines the
advantages of neutron activation analysis, by avoiding wet chemical extraction of
halogens (or K), with very high internal precision of *5% (2r).
Finally, the external precision of all methods is currently limited by the avail-
ability of well characterised standards for Br and I (Kendrick 2012; Kendrick et al.
2013a; Marks et al. 2017). In addition, some differences between laboratories could
be introduced by sample washing procedures, which might completely remove
water soluble components from powdered samples. Water soluble halogens are
especially important in marine sediments (Turekian and Wedepohl 1961) and
serpentinites (e.g., Sharp and Barnes 2004).
9.6.1 Mineralogy and Cl Content of Oceanic Crust
The altered oceanic crust comprises relict glass and nominally anhydrous minerals
typical of basalts and gabbros (e.g., plagioclase, pyroxene and olivine) and their
hydrous alteration products. Parts of the crust that have interacted with cold sea-
water contain clay minerals (smectites including montmorillonite, and saponite),
zeolites, and Fe-hydroxides. Glassy rocks present in hyaloclastites in Layer 2a of
the crust can be extensively replaced by palagonite, which is dened as a
heterogenous mixture of clays, zeolite, and oxides (Staudigel et al. 2008; Staudigel
and Hart 1983; Stroncik and Schmincke 2002). At temperatures of 200300 °C,
prehnite, pumpellyite, and chlorite are important. Actinolite, tremolite, albite, epi-
dote, and sphene become important at >300 °C, and hornblende and phlogopite can
be important at >400 °C (Alt and Honnorez 1984; Bideau et al. 1991; Talbi et al.
1999). In addition, a number of minerals, including carbonate and quartz, form over
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 621
wide temperature ranges. The polymorphs of serpentine (chrysotile, lizardite, and
antigorite) form in olivine-rich rocks between 100 and 500 °C (Bideau et al. 1991)
and talc forms by silica metasomatism under all prograde conditions (Seyfried et al.
1988). Typical water contents and what is known about the halogen contents of
some common alteration minerals are summarised in Table 9.4.
9.6.2 Low Temperature Seawater Alteration
Clay minerals contain a lot of water and signicant F, but typically very little Cl
(Table 9.4). For example smectites contain 515 wt% H
2
O and the smectite veins
analysed by Magenheim et al. (1995) had 150400 ppm F but only *20 ppm Cl.
Kendrick et al. (2015a) reported that palagonite crusts on Samoan glass contain
35 ppm Cl which represented a 97% depletion compared to the pristine glass
(Fig. 9.12). Although it is usually assumed that Cl is introduced into the oceanic
crust during hydrothermal alteration, the very low Cl content of these clay minerals
and palagonite demonstrates that parts of the oceanic crust could be depleted in Cl
Table 9.4 Halogens and water in selected minerals of the altered oceanic lithosphere
Mineral H
2
O wt.% Cl ppm F ppm Br ppm I ppb
Low to intermediate temperature
a
Clay minerals 515 2123 150400 (adsorbed I?)
Palagonite 4050 10s <20 5020,000
Zeolites 920 (adsorbed I?)
Fe-hydroxides 2540 >40008000 (?) D
Br/Cl
<1
High temperature
b
Amphibole 2 <10060,000 <10070,000 <0.52<570
Wide temperature range
c
Carbonate 1000 ±300 (?) <20,000 ?
Talc 59 400900 (?)
Serpentine 1214 10010,000 1100 124 5045,000
Fluorapatite 01 10002000 10,00030,000
Fluid inclusionscontribution to bulk Cl concentration
1 vol.% 0.35 70200 0.30.6 <120
References
a
Kendrick et al. (2015a); Magenheim et al. (1995); Seyfried et al. (1986); Stroncik and Schmincke
(2002)
b
Bideau et al. (1991); Cortesogno et al.(2004); Gillis and Meyer (2001); Ito and Anderson (1983);
Jacobson (1975); Nehlig and Juteau (1988); Prichard and Cann (1982); Vanko (1986); Kendrick
et al. (2015b)
c
Barnes et al. (2009); Bonifacie et al. (2008); Claret et al. (2010); Debret et al. (2014); John et al.
(2011); Kendrick et al. (2013b); Sharp and Barnes (2004)
622 M.A. Kendrick
during low temperature alteration. Alternatively, halogens might be remobilised on
a local scale (Floyd and Fuge 1982), with for example, Cl released during clay
alteration taken up by Fe-hydroxy chlorides (Table 9.4; Seyfried et al. 1986).
In comparison to Cl and F, even fewer data are available for Br or I. The
palagonite investigated by Kendrick et al. (2015a) was less depleted in Br, than Cl,
and it was enriched in I relative to pristine glass. Chavrit et al. (2016) also
demonstrated I enrichment in some clay-bearing samples of altered basalt.
Therefore parts of the crust could be enriched in I relative to unaltered crust, both
relative to Cl and in absolute terms (e.g., high I and high I/Cl). Iodine could be
trapped in altered ocean crust preferentially relative to Cl if it is adsorbed onto clay
minerals (Claret et al. 2010; Montavon et al. 2014) or zeolites, which are
micro-porous adsorbants. Alternatively, I is a biophilic element and there is
growing evidence that microbes are involved in some styles of glass alteration (Alt
and Mata 2000; Fisk et al. 1998; Kruber et al. 2008; Staudigel et al. 2008; Stroncik
and Schmincke 2002) and alteration of some minerals (Ivarsson et al. 2008). Kruber
et al. (2008) reported that palagonite formed by the bio-alteration of glass can
contain up to 1 wt% organic C (d
13
C=22). If the microbes responsible for this
alteration have an I/C
organic
similar to plankton (0.00020.002; Bobrov et al. 2005;
Iwamoto and Shiraiwa 2012; Martin et al. 1993; Price and Calvert 1977) this would
imply that exceptionally high I concentrations of *220 ppm might be expected in
biogenic palagonite.
In summary, low temperature alteration of the crust leading to hydration and
formation of clays causes a nite increase in uid salinity and some remobilisation
of heavy halogens from the crust. However, because this alteration takes place at
very high water-rock ratios the effects are difcult to discern in alteration uids
(Sect. 5.4). Low temperature alteration of mac lithologies, either by biological
processes or by adsorbtion of I onto reactive minerals, might be signicant for the
marine I cycle (cf., Leblanc et al. 2006).
9.6.3 High Temperature Hydrothermal Alteration
Amphibole and serpentine are usually the only minerals in which Cl (and more
rarely F) can be detected by electron microprobe. Other hydrous minerals such as
prehnite, pumpellyite, chlorite and epidote are assumed to contain <100 ppm Cl but
remove water from the system implying that they cause a nite increase in the
salinity of alteration uids. In addition to mineral matrices, a number of minerals
can be important hosts of uid inclusions. Fluid inclusions are often estimated to
account for up to 1 vol.% of coarse minerals with abundant inclusions. If these uid
inclusions trapped uids with average salinities and densities similar to seawater,
the uid inclusions can be estimated to contribute up to 70 ppm Cl and 250 ppb Br
but negligible F or I toward the minerals bulk composition (Table 9.4). In com-
parison, Kendrick et al. (2015b) reported bulk halogen concentrations of
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 623
100200 ppm Cl, 300600 ppb Br and 420 ppb I for uid inclusions bearing vein
minerals from the Mathematician Ridge.
The concentrations of F and Cl in amphiboles within the oceanic crust are
extremely variable ranging from concentrations of 10s of ppm Cl and 100s ppm F
in igneous amphibole up to wt% levels of Cl in late-stage hydrothermal veins
(Fig. 9.17; Cortesogno et al. 2004; Gillis and Meyer 2001; Jacobson 1975; Nehlig
and Juteau 1988; Vanko 1986). Actinolite, present under greenschist facies con-
ditions, generally contains <500 ppm Cl (Vanko 1986), which is consistent with
the concentration range of 100300 ppm expected to result from hydration by uids
with seawater salinity (Cortesogno et al. 2004; Kendrick et al. 2015b). In contrast,
amphibolite grade hornblendes contain hundreds to thousands of ppm Cl
(Cortesogno et al. 2004; Ito and Anderson 1983; Nehlig and Juteau 1988; Prichard
and Cann 1982; Tribuzio et al. 2014; Vanko 1986; Silantyev et al. 2008). The
variability of amphibole Cl concentrations under amphibolite conditions could
reect equilibration with segregated vapours and brines produced by phase sepa-
ration of seawater that is heterogeneously distributed through this part of the crust
(Cortesogno et al. 2004). Amphiboles with 1 to 6 wt% Cl have been reported from
several locations but are always a volumetrically minor component of the rock,
which is commonly associated with late-stage veins and/or mylonites (Bideau et al.
1991; Honnorez and Kirst 1975; Jacobson 1975; Vanko 1986). The effect of
hydrothermal alteration is to introduce seawater Cl and redistribute igneous Cl and
F between a growing number of amphiboles. As a result, amphiboles formed during
progressive hydrothermal alteration are distinguished from igneous amphiboles by
progressively higher Cl contents and Cl/Na ratios, and lower F/Cl ratios (Fig. 9.17;
Coogan et al. 2001; Mevel 1988).
Crystal chemistry exerts an important control on amphibole halogen concen-
tration. For example, the negative correlation between F/Cl and Cl in Fig. 9.17
results in part from F and Cl competing to occupy the same site in amphibole. It has
long been assumed that the large Br
and I
anions would be preferentially
Fig. 9.17 F/Cl versus Cl
concentration for amphiboles
in the oceanic crust. The data
of Coogan and Cullen (2009)
and Cortesogno et al. (2004)
were obtained by ion
microprobe, whereas the data
of Gillis and Meyer (2001)
and Vanko (1986) are based
on electron microprobe
analyses
624 M.A. Kendrick
excluded from amphibole relative to Cl
and F
(Svensen et al. 1999,2001), and
this has recently been conrmed by the analysis of amphibole and related uid
inclusions in amphibolites and metagabbros from the Mathematician Ridge
(Kendrick et al. 2015b).
In addition to crystal chemistry, it is now generally agreed that high Cl am-
phiboles can only form when the chemical activity of Cl is high and the chemical
activity of H
2
O is low, such as occurs in the presence of saline uids (Kullerud and
Erambert 1999; Markl and Bucher 1998; Vanko 1986). Therefore, Cl-rich amphi-
boles do not provide a mechanism for reducing uid salinity; rather, they provide
evidence for the presence of high salinity uids in the crust.
If phase separation is the major control on uid salinity (Kelley and Delaney
1987; Vanko 1988), the increase in amphibole Cl content from greenschist to
amphibolite facies can be interpreted as evidence for efcient segregation of brines
and vapours and support for the idea that brines are preferentially stored in the
deeper crust (e.g., Fig. 9.14). However, while it has been recognised that crustal
hydration can lead to appreciable increases in uid salinity, the effect of drying up
or uid desiccationon salinity has not been evaluated within the context of the
oceanic crust. For example, Kelley and Delaney (1987) suggest that formation of
hydrous minerals could increase seawater salinity by a factor of two but favoured
phase separation in the generation of ultra-saline brines. In contrast to this, it is well
known that at water/rock ratios of less than 0.10.01 drying upleads to substantial
increases in the concentration of all dissolved components (Reed 1997).
Furthermore, desiccation of metamorphic uids is implicated in the generation of
Cl-rich amphiboles and metamorphic salt in granulites in northern Norway, which
provides strong evidence for the generation of ultra-saline uids (up to 100% salt)
by this process in some settings (Kullerud and Erambert 1999; Markl and Bucher
1998). The high Cl content of amphibolite facies amphiboles can therefore also be
interpreted as indicating uid/rock ratios approach zero at the limit of the seaoor
hydrothermal system (cf., Fig. 9.14).
The relative importance of phase separation and drying upmight be tested in
future studies of Cl, Br, and I in uid inclusions and minerals. As Br and I are
excluded from amphibole relative to Cl (Kendrick et al. 2015b), drying up should
produce uids with very high Br/Cl and I/Cl ratios (Svensen et al. 2001). Whereas
because Br/Cl does not appear to be strongly fractionated by phase separation
(Fig. 9.11), phase separation might yield brines with Br/Cl and I/Cl much closer to
seawater values (Kendrick et al. 2015b). This line of reasoning suggests that phase
separation was the dominant process responsible for the generation of brines in the
Mathematician Ridge hydrothermal system (Kendrick et al. 2015b) and brines
assimilated by magmas which are all inferred to have Br/Cl ratios of only slightly
higher than seawater (Fig. 9.15; Kendrick et al. 2013a,2015a,2017).
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 625
9.6.4 The Bulk Halogen Content of the Oceanic Crust
Sano et al. (2008) sampled core from Hole 1256D drilled in the Eastern Pacicat
intervals of <50 m, over more than 1000 m of core, providing high density Cl
concentration data that show a broad increase in Cl as a function of depth in the
oceanic basement sampled by this hole (Fig. 9.18). Four additional studies have
investigated Cl and Cl isotopes in seven different cores drilled in the Pacic,
Atlantic, and Indian Oceans but sampling within each core is at a much lower
density (Barnes and Cisneros 2012; Bonifacie et al. 2007; Chavrit et al. 2016;
Magenheim et al. 1995). These studies show similar Cl concentration ranges as
observed in Hole 1256D, in which low temperature alteration has a median Cl
concentration of 160 ppm and high temperature alteration has a median concen-
tration of 360 ppm Cl, but individual samples contain as little as 11 ppm or as
much as 2000 ppm Cl (Fig. 9.18; Barnes and Cisneros 2012; Bonifacie et al. 2007;
Chavrit et al. 2016; Magenheim et al. 1995; Sano et al. 2008). These studies
indicate that the real concentration of Cl in the oceanic crust is much higher than
was initially estimated (50 ±25 ppm Cl) on the basis of electron microprobe data
(Ito et al. 1983).
Unfortunately, because crustal lithologies have extremely variable Cl concen-
trations (e.g., 2001200 ppm Cl at 1500 m in Hole 1256D; Fig. 9.18), the low
density of data available from the majority of IODP holes means that it is difcult to
assess if the broad increase in Cl observed with depth in Hole 1256D (Fig. 9.18;
Sano et al. 2008) is a general feature of the oceanic crust (cf., Barnes and Cisneros
2012). However, the following observations suggest that Cl concentrations are
Fig. 9.18 The concentrations of aCl and bK and cthe K/Cl ratio of whole rock samples from
IODP Hole 1256D in the East Pacic is shown as a function of depth below the seaoor. The data
are from Sano et al. (2008) and Barnes and Cisneros (2012)(green dots in part a). The alteration
mineralogy shown on the gure is simplied. The median Cl and K concentrations of low and high
temperature alteration zones, with 2runcertainties, are shown for comparison
626 M.A. Kendrick
likely to increase with depth: (i) Cl and Br are mobilised into the uid during
alteration of glass (Fig. 9.12; Floyd and Fuge 1982; Kendrick et al. 2015a); (ii) low
temperature alteration minerals have very low Cl concentrations (Sect. 9.6.2;
Table 9.4); and (iii) amphibole Cl concentrations increase from greenschist- to
amphibolite-facies providing evidence for uids becoming increasingly saline at
depth in the oceanic crust (Sect. 9.6.3).
Finally, the concentration range of Cl in pristine mid-ocean ridge glasses is
contrasted with the concentration range in altered oceanic rocks in Fig. 9.19. The
pristine glasses include N-MORB and E-MORB and are variably evolved, with
MgO mainly between 4 and 9 wt% but including some dacites from the Galapagos
Spreading Centre with 1 wt% MgO. The median Cl concentration of these glasses
(120 ±20 ppm) is suggested here as a proxy for the initialconcentration of Cl in
layer 2 (lavas and dykes) of the oceanic crust. In contrast, cumulate minerals in
layer 3 gabbros will have a much lower initial concentration of Cl (and other
incompatible elements) implying much lower initial Cl concentrations in deeper
Fig. 9.19 The Cl
concentrations of pristine
MORB glasses and altered
ocean crust. The data dene
log-normal distributions that
are typical of trace elements.
The AOC is indicated to have
twice the Cl content of
pristine crust on average.
Glass data include N-MORB,
E-MORB, and evolved
glasses from the Galapagos
spreading centre that have
assimilated seawater Cl
(Danyushevsky et al. 2000;
Kamenetsky and Eggins
2012; Kendrick et al. 2012a,
2013a; le Roux et al. 2002;
Michael and Cornell 1998;
Saal et al. 2002; Sims et al.
2002,2003). Altered Ocean
Crust data are from Barnes
and Cisneros (2012),
Bonifacie et al. (2007), Floyd
and Fuge (1982), Magenheim
et al. (1995), and Sano et al.
(2008)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 627
portions of the crust. The suggested initial Cl concentration of layer 2
(120 ±20 ppm Cl) is just within uncertainty of the median obtained for the upper
portion of Hole 1256D (160 þ140
50 ppm Cl; Figs. 9.18 and 9.19). Considering the
high ux of seawater and very high water/rock ratios in this part of the crust, the
lack of pronounced Cl enrichment demonstrates that the dominant effect of low
temperature alteration is probably halogen remobilisation and exchange of halogens
between seawater and the crust, rather than halogen enrichment. In contrast, the
median concentration of Cl in the lower portion of Hole 1256D (360 þ120
60 ppm Cl)
is probably >330 times higher than the initial concentration of Cl in the gabbro
cumulates (Fig. 9.18). Therefore, it appears that hydrothermal alteration may
mobilise Cl in the upper crust and strongly enrich Cl in the lower crust (Figs. 9.18
and 9.19). In contrast to Cl, K exhibits the opposite behaviour, being most enriched
during low temperature alteration in the upper crust. Therefore, the K/Cl ratio,
which is commonly used to measure relative Cl enrichment, changes with
increasing crustal depth and alteration much more sharply than the absolute
abundance of either element (Fig. 9.18c).
9.6.5 Halogens in Serpentinites
The halogen content of seaoor serpentinites formed at varying distances from the
mid-ocean ridges, including forearc settings, has been investigated in several
studies (Barnes et al. 2009; Barnes and Sharp 2006; Bonifacie et al. 2008; Boschi
Fig. 9.20 Histogram and
probability density function of
serpentinite Cl concentration
data. As with all trace
elements, Cl tends toward a
log-normal distribution,
meaning the median is better
dened than the mean. Data
from Barnes et al. (2006,
2009), Bonifacie et al. (2008),
Debret et al. (2014), John
et al. (2011), Kendrick et al.
(2013b), Kodolányi et al.
(2012), Scambelluri et al.
(2004) and Sharp and Barnes
(2004)
628 M.A. Kendrick
et al. 2013; Debret et al. 2014; John et al. 2011; Kendrick et al. 2013b; Kodolányi
et al. 2012; Scambelluri et al. 2004; Sharp and Barnes 2004). Taken together the
data show serpentinites contain from *100 ppm to 1 wt% Cl, with a median
concentration of 1500 þ300
200 ppm Cl (Fig. 9.20). The maximum concentration of Cl
in serpentine is somewhat less than amphibole, but serpentinite rocks, wholly
dominated by serpentine, have a higher H
2
O and Cl content than any other major
subduction zone lithology (Scambelluri et al. 1997; Schmidt and Poli 1998).
The I/Cl ratio of serpentinites is not correlated with Cl concentration (Fig. 9.21a);
however, serpentinites have combined Br/Cl and I/Cl that overlap the range of
sedimentary marine pore waters and the serpentinites formed at the greatest distances
from mid-ocean ridges have the highest I/Cl ratios (Fig. 9.22; Kendrick et al.
2013b). These data can be explained if sedimentary marine pore waters, as well as
seawater, are involved in serpentinisation (Kendrick et al. 2011b,2013b).
Fig. 9.21 Halogen
abundance ratios as a function
of Cl concentration (ppm) in
lizardite and chrysotile
pre-subduction
serpentinites. aI/Cl and
bF/Cl (antigorite
serpentinites are not included
in these plots). The data of
Kendrick et al. (2013b) are
differentiated the same way as
in Fig. 9.22. Additional data
are from John et al. (2011)
and Debret et al. (2014)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 629
Sedimentary marine pore waters could be particularly important during serpentini-
sation at the slab bend and in the forearc (Fig. 9.5; Ranero et al. 2003). The small
mismatch between serpentinite Br/Cl-I/Cl and sedimentary marine pore waters could
be explained in several ways. The high I and Br content of the serpentinites attests to
the high compatibility of all halogens in serpentine. However, it is likely that halogen
abundance ratios are fractionated between serpentinites and serpentinising uids to
some degree. Alternatively, if serpentinites preserve Br/Cl and I/Cl ratios close to
that of the serpentinising uids, the data indicate that in addition to seawater and
sediment-derived halogens, some serpentinites contain an additional small mantle
component that could be mobilised from igneous lithologies in the overlying crustal
rocks (Figs. 9.22 and 9.23). Slab uids with fractionated I/Cl are then required to
explain the highest I/Cl of 0.04 in forearc serpentinites because it is greater than
observed in sedimentary pore waters (Fig. 9.22a; Kendrick et al. 2013b).
In contrast to I/Cl and Br/Cl, the F/Cl ratio is negatively correlated with Cl in
serpentinites (Fig. 9.21b) in a similar manner to that observed for amphibole
(Fig. 9.17). It is likely that the compatibility of F is at least as high as the heavy
halogens in serpentine, but the negative relationship between F/Cl and Cl can be
explained if serpentinising uids introduce variable quantities of seawater and
sediment-derived Cl, Br, and I, and relatively little F (Figs. 9.22 and 9.23). There
Fig. 9.22 Br/Cl versus I/Cl for serpentinites: alinear scale and blog-log scale. The trend of
sedimentary marine pore waters (Fig. 9.3), gas hydrate, MORB, and the compositions on
amphibole and mica are shown for reference. Serpentinite data from Kendrick et al. (2011b,
2013b); MORB eld from Kendrick et al. (2013a,2017); mica/amphibole data from Kendrick
(2012) and Kendrick et al. (2015b), see Fig. 9.15
630 M.A. Kendrick
are very few combined F/Cl and I/Cl data available for sedimentary pore waters
(Mahn and Gieskes 2001). However, the low solubility of F in sedimentary pore
waters (Frohlich et al. 1983; Rude and Aller 1994) suggests F/Cl is likely to be
maintained fairly close to the seawater value. Therefore, the higher F/Cl of ser-
pentinites and some hydrothermal uids, relative to seawater, probably results from
mobilisation of F from igneous lithologies in the crust (Fig. 9.23).
The median Cl concentration of 1500 þ300
200 ppm in serpentinites corresponds to a
H
2
O/Cl weight ratio of *80 that is greater than the seawater H
2
O/Cl value of 50.
Serpentinites with much higher Cl concentrations are probably produced as the
water drys upand the salinity of the serpentinising uid increases (e.g., H
2
O/Cl
decreases). Kendrick et al. (2011a,b,2013b) suggested a water/rock ratio of <0.1
and that the uid being spent during serpentinisation (e.g., dried up) might be
required to explain the preservation of sediment pore water halogen signatures
(Fig. 9.22) and high noble gas concentrations in serpentine. However, metamorphic
salts produced by complete drying up (e.g., Markl and Bucher 1998) have not been
identied in crushed serpentinites. The nature of the water soluble Cl component
in serpentinites is still unclear (Kodolányi et al. 2012) and experimental studies are
required to determine halogen partitioning in serpentinites.
9.7 Synthesis and Future Directions
As shown above there are a relatively large number of data available for F, Cl, Br,
and I in sedimentary pore waters and hydrothermal uids, but the data base for
multiple halogens in minerals within sediments and the crustal basement is much
Fig. 9.23 Serpentinite F/Cl
versus I/Cl data. Serpentinite
data from John et al. (2011).
Pore uid data from Gieskes
and Mahn (2007). Escanaba
Trough high
3
He/
4
He pore
uid data (representative of
hydrothermal uids) from
Gieskes et al. (2002)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 631
smaller. As a result, while Cl, Br, and I are known to exhibit systematic behaviour
in sedimentary pore waters and the organic phase within unconsolidated sediment
(Figs. 9.2 and 9.3), the relative abundance ratios of F/Cl, Br/Cl, and I/Cl and the
halogen concentrations in lithied sediments are less well known. Chlorine and F
data, and limited Br and I data, are available for altered basalts and gabbros, and
serpentinites. Despite the limitations imposed by data availability, some general
statements about the most important lithologies and guestimates of likely concen-
tration ranges in key crustal lithologies are possible. This is undertaken below for
each of the halogens in a representative crustal sectionentering a subduction zone.
This is undertaken for the crust immediately prior to subduction, because the nal
state of the oceanic crust is relevant for assessing possible uxes of halogens into
the deep mantle and through magmatic arcs, which is a critical parameter given the
role of halogens in complexing metals and facilitating trace element transport in
slab-derived uids (Migdisov and Williams-Jones 2014; Yardley 2005).
The representative crustal section investigated comprises 400 m of marine
sediment, 1500 m of layer 2 lavas and dykes, 4500 m of layer 3 gabbros, and
500 m of serpentinites (Fig. 9.24). These thicknesses are based on average crust
(Chen 1992; Snow and Edmonds 2007) and a band of serpentinite, which could
represent either intense local zones of serpentinisation, or diffuse serpentinisation of
Fig. 9.24 Schematic diagram of a representative, oceanic lithospheric cross section with
estimated Cl concentrations at the point of subduction. Note that the average extent of alteration
in the deep crust and lithospheric mantle are poorly dened and these gures assume signicant
alteration at the slab bend prior to subduction (see text for discussion)
632 M.A. Kendrick
the uppermost 10 km of the lithospheric mantle formed at the slab bend (Fig. 9.5).
The degree of lithospheric serpentinisation is still poorly known, and given that
additional serpentinisation occurs in the forearc (Bostock et al. 2002), this gure is
considered a conservative lower limit for the amount of serpentinite subducted into
the mantle.
The existing data suggest that Cl concentrations increase with depth in the
altered oceanic crust (Fig. 9.18). However, for simplicity we assume that the upper
1000 m of the oceanic crust, which has been subjected to relatively low temperature
alteration, has a median concentration of 160 þ140
50 ppm Cl, and that the lower
5000 m has a median concentration of 360 þ120
60 ppm Cl (cf., Figs. 9.18 and 9.24). It
is possible that away from the subduction zones a large portion of the crust, at
depths of 46 km, could be unaltered. However, it is assumed that at the point of
subduction, uids have penetrated the entire crustal section, which is consistent
with seismic evidence for serpentinisation of the mantle lithosphere at the slab bend
(Fig. 9.5; Ranero et al. 2003) and heat ow measurements (Grevemeyer et al.
2005). In this scenario, it is likely that at increasing depths, zones of relatively
unaltered crust are transected by increasingly Cl-rich zones of alteration, reecting
channelized uid ow within deep fracture systems. This would lead to extremely
variable concentrations of Cl in the deeper crust, similar to that observed in existing
drill hole data and dredge samples; for example, Fig. 9.18 shows concentrations
ranging from 200 to 1200 ppm Cl at *1200 m in Hole 1256D (Sano et al. 2008).
In comparison to the oceanic crust, sediments are assumed to have a poorly dened
Cl concentration of 700 ±500 ppm (cf., Table 9.2; Muramatsu and Wedepohl
1998) and serpentinites are assigned the median value of 1500 þ300
200 ppm Cl
(Figs. 9.20 and 9.24). These gures suggest that the subducting lithosphere could
have a bulk Cl concentration on the order of 400 ppm (Fig. 9.24). This gure is
close to the median value suggested for altered ocean crust (Fig. 9.24), and implies
that due to its size, the altered ocean crust could be the largest single Cl reservoir in
subducting slabs (Fig. 9.25). However, it should be stressed that there is consid-
erable uncertainty concerning the degree of deep-crust and lithosphere alteration,
and that the relative sizes of the altered ocean crust and serpentinite reservoirs in
Fig. 9.24 could be substantially different.
Fig. 9.25 Pie charts showing the relative importance of sediments, altered ocean crust and
serpentinites as reservoirs for F, Cl, Br and I in the oceanic lithosphere, under the assumptions
used to construct Fig. 9.24 (Table 9.5; Sect. 9.7)
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 633
In contrast to Cl, there are fewer data available for F. However, because F is
mobilised over relatively short distances during alteration and is not signicantly
introduced by seawater, we assume the F concentration in layer 2 of altered ocean
crust is similar to that of MORB. The average F concentration of MORB is esti-
mated as 260 ±160 ppm F, based on a F/Nd = 20 ±12 (Workman et al. 2006)
and Nd of *13 ppm, which is 10 times the primitive mantle value (Hofmann
2003). Serpentinites have typical F concentrations of 1100 ppm (Debret et al.
2014; John et al. 2011), which overlap depleted mantle values of 1116 ppm F
(Pyle and Mather 2009). We assume that where a F enrichment is observed, the F
has been mobilised from overlying lithologies in layers 23 of the crust (Fig. 9.23).
Marine sediments are assigned a concentration of 1000 ±300 ppm F (Fig. 9.1;Li
1991; Rude and Aller 1991). These assumptions suggest that taken as a whole the
altered oceanic lithosphere contains on the order of 100 ±50 ppm F (Table 9.5).
Therefore, the altered oceanic lithosphere has a F/Cl weight ratio of *0.25 com-
pared to *2 for median MORB, with the difference resulting from up to *90% of
crustal Cl being introduced by hydrothermal alteration.
The Br and I content of altered oceanic crust is poorly known. It has been
suggested that low temperature clay-rich alteration could be enriched in I and have
higher I/Cl than MORB, whereas high temperature amphibole alteration has Br/Cl
and I/Cl of less than MORB (Sects. 9.6.2 and 9.6.3; Chavrit et al. 2016; Kendrick
et al. 2015b). If the altered oceanic crust retains Br/Cl and I/Cl within a factor of 2
of MORB, we can estimate it has ppb-levels of Br and I. In contrast, sediments and
serpentinites both have extremely variable Br/Cl and I/Cl ratios with ppm levels of
Br and I (Tables 9.2 and 9.5). In contrast to the altered oceanic crust, serpentinites
formed in the mantle lithosphere can incorporate substantial Br and I and may even
preserve the Br/Cl, I/Cl, and F/Cl ratios of the serpentinising uids (cf., Figs. 9.22
and 9.23). The small amount of data available mean that the average Br/Cl and I/Cl
of the serpentinites is not known (Fig. 9.22). However, even under the conservative
assumptions made above regarding the extent of lithospheric serpentinisation
Table 9.5 Summary of halogens in sediments, altered ocean crust, and serpentinites
Thickness
(km)
F
ppm
Cl
ppm
Br
ppm
I
ppm
Marine
sediments
0.4 1000 ±300 700 ±500 60 ±60 30 ±25
Mature/lith.
seds
1000 ±300 700 ±500 12 ±10 4 ±3
AOClayer 2 11.5 260 ±160 160 þ140
50 0:3þ0:3
0:20:02 þ0:02
0:01
AOClayer 3 4.5550±40 360 þ120
60 0.7 ±0.4 0.01 ±0.01
Serpentinites 0.5 50 ±40 1500 þ300
200 5±1 1.5 ±1.5
Bulk
lithosphere
6.9 100 ±50 380 þ140
90 3±2 1.6 ±1.5
Data from Tables 9.2 and 9.4 and Figs. 9.1,9.2,9.3,9.4,9.5,9.6,9.7,9.8,9.9,9.10,9.11,9.12,
9.13,9.14,9.15,9.16,9.17,9.18,9.19,9.20,9.21,9.22 and 9.23.AOC Altered Ocean Crust
634 M.A. Kendrick
(Fig. 9.24), serpentinite rocks rich in chrysotile or lizardite have the potential to be
a major reservoir of Br and I (Table 9.5; Fig. 9.25). As a result, serpentinites have
the potential to be an important pathway for Br and I into subduction zones,
however, it has been suggested that Br and I are preferentially lost relative to Cl
during high grade metamorphism and antigoritisation of serpentine meaning that Br
and I are unlikely to be strongly subducted into the deep mantle (cf., Fig. 9.25; John
et al. 2011; Kendrick et al. 2011b,2013b,2014,2017).
Further work combining the analysis of multiple halogens in hydrothermal uids
is now desirable. In particular, analysis of I in hydrothermal uids together with Br
and Cl would help conrm if hydrothermal uids with elevated Br/Cl ratios have
interacted with sediments (cf., Fig. 9.11). In addition, it is important to improve
linkages in understanding between the behaviour of halogens in hydrothermal uids
and changes in alteration mineralogy. This is required to test the range of conditions
under which Br and Cl can be safely regarded as behaving conservatively. The
possible importance of low temperature alteration for the marine I cycle is yet to be
evaluated and further data are required for all the halogens in bulk rock samples,
individual mineral phases, and uid inclusions within the oceanic crust and litho-
spheric mantle. The extreme heterogeneity of alteration in the oceanic crust and
lithosphere implies a high density of data is required to identify broad trends in
changing halogen concentrations laterally or as a function of depth (e.g., Fig. 9.18).
Acknowledgements Mark A. Kendrick is supported by an Australian Research Council Future
Fellowship (FT130100141). I am grateful to Zach Sharp and Michael Mottl for their constructive
reviews of this chapter and also to several colleagues whom I have collaborated with, or discussed
halogen geochemistry, over the years.
References
Alt J, Honnorez J (1984) Alteration of the upper oceanic crust, DSDP site 417: mineralogy and
chemistry. Contrib Mineral Petrol 87:149169
Alt JC, Mata P (2000) On the role of microbes in the alteration of submarine basaltic glass: a TEM
study. Earth Planet Sci Lett 181:301313
Aranovich LY, Bortnikov NS, Serebryakov NS, Sharkov EV (2010) Conditions of the formation
of plagiogranite from the Markov trough, Mid-Atlantic Ridge, 5A degrees 52-6A degrees 02
N. Dokl Earth Sci 434(1):12571262
Aranovich LYa, Prokoev VYu, Pertsev AN, Bortnikov NS, Ageeva OA, Beltenev VE,
Borisovsky SE, Simakin SG (2015) Composition and origin of a K
2
O-rich granite melt in the
mid-Atlantic Ridge, 13°34N: evidence from the analysis of melt inclusions and minerals of
the gabbroplagiogranite association. Dokl Earth Sci 460:174178
Anderson BW, Coogan LA, Gillis KM (2012) The role of outcrop-to-outcrop uid ow in off-axis
oceanic hydrothermal systems under abyssal sedimentation conditions. J Geophys Res Solid
Earth 117(B05):103
Anderson RN, Zoback MD, Hickman SH, Newmark RL (1985) Permeability versus depth in the
upper oceanic crustin situ measurements in DSDP Hole 504B, Eastern Equatorial Pacic.
Earth Planet Sci Lett 90:36593669
9 Halogens in Seawater, Marine Sediments and the Altered Oceanic 635
Baker ET (2007) Hydrothermal cooling of midocean ridge axes: do measured and modeled heat
uxes agree? Earth Planet Sci Lett 263:140150
Baker ET, German CR (2004) On the global distribution of hydrothermal vent elds. In:
German CR, Lin J, Parson LM (eds) Mid-ocean ridges: hydrothermal interactions between the
lithosphere and oceans. American Geophysical Union, Washington, pp 245266
Barnes JD, Cisneros M (2012) Mineralogical control on the chlorine isotope composition of
altered oceanic crust. Chem Geol 326327:5160
Barnes JD, Sharp ZD (2006) A chlorine isotope study of DSDP/ODP serpentinized ultramac
rocks: Insights into the serpentinization process. Chem Geol 228:246265
Barnes JD, Paulick H, Sharp ZD, Bach W, Beaudoin G (2009) Stable isotope (d
18
O, dD, d
37
Cl)
evidence for multiple uid histories in mid-Atlantic abyssal peridotites (ODP Leg 209). Lithos
110:8394
Baturin GN (1988) The geochemistry of manganese and manganese nodules in the ocean.
D. Reidel Publishing Company, Boston, 342 pp
Baturin GN, Dubinchuk VT (2011) Mineralogy and chemistry of ferromanganese crusts from the
Atlantic Ocean. Geochem Int 49:578593
Bein A, Hovorka SD, Fisher RS, Roedder E (1991) Fluid inclusions in bedded permian halite, Palo
Duro Basin, TexasEvidence for modication of seawater in evaporite brine-pools and
subsequent early diagenesis. J Sediment Petrol 61:114
Berndt ME, Seyfried WE (1990) Boron, bromine and other trace-elements as clues to the fate of
chlorine in midocean ridge vent uids. Geochim Cosmochim Acta 54:22352245
Berndt ME, Seyfried WE (1997) Calibration of Br/Cl fractionation during subcritical phase
separation of seawater: possible halite at 9 to 10° N East Pacic Rise. Geochim Cosmochim
Acta 61:28492854
Berndt ME, Seyfried WE Jr, Janecky DR (1989) Plagioclase and epidote buffering of cation ratios
in mid-ocean ridge hydrothermal uids: experimental results in and near the supercritical
region. Geochim Cosmochim Acta 53:22832300
Bernini D, Wiedenbeck M, Dolejs D, Keppler H (2013) Partitioning of halogens between mantle
minerals and aqueous uids: implications for the uid ow regime in subduct