ArticlePDF Available

Characterization of hydrothermal fluids that alter the upper oceanic crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites

Authors:

Abstract

Pervasive alteration of basaltic oceanic crust by heated seawater at greenschist facies conditions produces two contrasting hydrothermal rocks. “Spilites”, consisting of chlorite + albite + quartz ± actinolite ± epidote, occur typically with regional extents. Locally spilites are metasomatically transformed to “epidosites” consisting of epidote + quartz + titanite + hematite or magnetite. Both alteration types have been proposed as markers of deep hydrothermal upflow in sub-seafloor convection cells, and as sources of the ore metals in basalt-hosted seafloor massive sulfide deposits. Little direct evidence is available for the chemical compositions of these fluids in their states deep in the upflow zones prior to their discharge at the seafloor. To better characterize them we have conducted a field, petrographic and fluid inclusion study of the lavas, sheeted dikes and plagiogranites in the Semail ophiolite, with supporting samples from the Troodos ophiolite. Our results show that both the spilite- and epidosite-forming fluids were single-phase aqueous liquids during the hydrothermal alteration. At some sites their salinity is 3.1–3.2 wt.% NaCleq, which we take to represent the chlorinity of Cenomanian seawater in the Semail realm. At other sites salinities are as low as 2.4 wt.% NaCleq or as high as 5.7 wt.% NaCleq, attributable to liquid–vapor separation and partial remixing deep in the crust along the dew curve of seawater, prior to ascent of the fluids to the sites of fluid inclusion trapping. Hypersaline brines, often accompanied by vapor, are restricted to plagiogranites in both the Semail and Troodos ophiolites and they represent magmatic–hydrothermal fluids that pre-date and are genetically unrelated to the spilite and epidosite alteration. The volcanostratigraphic locations of the samples constrain their maximum depths to 1470–3600 m below seafloor during alteration. The range of possible fluid trapping pressures for all samples is 31–68 MPa. Trapping temperatures vary between sites from 145 to 440 °C for spilite fluids and 255 to 435 °C for epidosite fluids. Quantitative analyses of 12 elements in individual fluid inclusions by LA-ICP-MS define the chemical characters of the two alteration fluids. The Br/Cl ratio in the spilite fluid is the same as in modern seawater and the other elements fit expectations from seawater–basalt experiments at elevated temperature. Accordingly, concentrations of Li, B, Na, Cl, K, Br and Sr in the spilite fluid match those in modern black-smoker vent fluids in basaltic crust. Exceptions are Ca and Fe, which are enriched in the spilite fluid. As these elements may precipitate below or at the seafloor prior to vent sampling, we conclude that the spilite fluids are plausible feeders of basalt-hosted black-smoker vents. The epidosite fluid has broadly similar elemental concentrations to the spilite fluid, but vastly lower Fe, reflecting the highly oxidized state of epidosites. This suggests that epidosite fluids are incapabable of forming basalt-hosted seafloor massive sulfide deposits.
Characterization of hydrothermal fluids that alter the
upper oceanic crust to spilite and epidosite: Fluid
inclusion evidence from the Semail (Oman) and Troodos
(Cyprus) ophiolites
Lisa Richter
, Larryn W. Diamond
Institute of Geological Sciences, University of Bern, Baltzerstrasse 3, CH-3012 Bern, Switzerland
Received 10 January 2021; accepted in revised form 9 November 2021; available online xxxx
Abstract
Pervasive alteration of basaltic oceanic crust by heated seawater at greenschist facies conditions produces two contrasting
hydrothermal rocks. ‘‘Spilites, consisting of chlorite + albite + quartz ± actinolite ± epidote, occur typically with regional
extents. Locally spilites are metasomatically transformed to ‘‘epidositesconsisting of epidote + quartz + titanite + hematite
or magnetite. Both alteration types have been proposed as markers of deep hydrothermal upflow in sub-seafloor convection
cells, and as sources of the ore metals in basalt-hosted seafloor massive sulfide deposits. Little direct evidence is available for
the chemical compositions of these fluids in their states deep in the upflow zones prior to their discharge at the seafloor. To
better characterize them we have conducted a field, petrographic and fluid inclusion study of the lavas, sheeted dikes and pla-
giogranites in the Semail ophiolite, with supporting samples from the Troodos ophiolite. Our results show that both the
spilite- and epidosite-forming fluids were single-phase aqueous liquids during the hydrothermal alteration. At some sites their
salinity is 3.1–3.2 wt.% NaCl
eq
, which we take to represent the chlorinity of Cenomanian seawater in the Semail realm. At
other sites salinities are as low as 2.4 wt.% NaCl
eq
or as high as 5.7 wt.% NaCl
eq
, attributable to liquid–vapor separation
and partial remixing deep in the crust along the dew curve of seawater, prior to ascent of the fluids to the sites of fluid inclu-
sion trapping. Hypersaline brines, often accompanied by vapor, are restricted to plagiogranites in both the Semail and Troo-
dos ophiolites and they represent magmatic–hydrothermal fluids that pre-date and are genetically unrelated to the spilite and
epidosite alteration. The volcanostratigraphic locations of the samples constrain their maximum depths to 1470–3600 m below
seafloor during alteration. The range of possible fluid trapping pressures for all samples is 31–68 MPa. Trapping temperatures
vary between sites from 145 to 440 °C for spilite fluids and 255 to 435 °C for epidosite fluids. Quantitative analyses of 12
elements in individual fluid inclusions by LA-ICP-MS define the chemical characters of the two alteration fluids. The
Br/Cl ratio in the spilite fluid is the same as in modern seawater and the other elements fit expectations from seawater–basalt
experiments at elevated temperature. Accordingly, concentrations of Li, B, Na, Cl, K, Br and Sr in the spilite fluid match
those in modern black-smoker vent fluids in basaltic crust. Exceptions are Ca and Fe, which are enriched in the spilite fluid.
As these elements may precipitate below or at the seafloor prior to vent sampling, we conclude that the spilite fluids are
plausible feeders of basalt-hosted black-smoker vents. The epidosite fluid has broadly similar elemental concentrations to
the spilite fluid, but vastly lower Fe, reflecting the highly oxidized state of epidosites. This suggests that epidosite fluids are
incapabable of forming basalt-hosted seafloor massive sulfide deposits.
Ó2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/
licenses/by/4.0/).
Keywords: Spilite; Epidosite; Fluid inclusion; Hydrothermal alteration; Oceanic crust; Ophiolite; Massive sulfide; VMS
https://doi.org/10.1016/j.gca.2021.11.012
0016-7037/Ó2021 The Authors. Published by Elsevier Ltd.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Corresponding author.
E-mail address: lisa.richter@gmx.net (L. Richter).
www.elsevier.com/locate/gca
Available online at www.sciencedirect.com
ScienceDirect
Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
1. INTRODUCTION
Observations of in-situ oceanic crust, ophiolites and
Archean greenstone belts show that vast quantities of basal-
tic lavas, dikes and intrusive rocks have been pervasively
altered by deep circulation of magmatically heated seawater
(Gillis and Banerjee, 2000; Hannington et al., 2003;
Staudigel, 2014). The associated leaching and transport of
elements, which ultimately influences ocean chemistry and
forms volcanogenic massive sulphide (VMS) deposits, is
mostly ascribed to regions of the crust that are hydrother-
mally altered at greenschist facies P–T conditions. There,
the alteration mineralogy testifies to sufficiently high tem-
peratures to raise mineral solubilities (200–450 °C;
German and Von Damm, 2006; Hannington, 2014), while
the intensity and extent of alteration testifies to sufficiently
high water–rock ratios to effect significant exchanges of
mass.
Two contrasting types of alteration form under these
P–T conditions. Volumetrically dominant is ‘‘spilite
alteration (Cann, 1969), by which the basalts are trans-
formed to albite–chlorite–quartz ± epidote ± actinolite ±
titanite ± hematite or magnetite, often preserving relict
igneous clinopyroxene. The alteration enriches the rocks
in Na and Mg and depletes them in Ca (along with other
element transfers). This metasomatism is typical of the
sheeted dike complexes in the in-situ oceanic crust, a sub-
type of the ‘‘backgroundor rock-matrix alteration
described by Heft et al. (2008) and Alt et al. (2010). The sec-
ond type of alteration involves pervasive transformation of
pre-existing spilites into massive ‘‘epidosites. These consist
essentially of epidote + quartz with minor
titanite ± hematite or magnetite in their end-member state,
although most epidosite bodies contain relict spilitic chlo-
rite and igneous clinopyroxene (Richardson et al., 1987;
Schiffman and Smith, 1988; Harper, 1995; Weber et al.,
2021). The epidosite mineralogy reflects a reversal of the
spilite metasomatism: Ca is enriched in the rock while Mg
and Na are quantitatively depleted (along with other ele-
ment transfers). Quartz–epidote veins are locally associated
with massive epidosites. In the in-situ oceanic crust, quartz–
epidote veins are widespread but massive replacement epi-
dosites have been found only rarely (Quon and Ehlers,
1963;Vanko et al., 1992;Banerjee et al., 2000; Teagle
et al., 2006). However, massive epidosites are common in
basaltic lavas in Archean greenstone belts (Galley, 1993;
Hannington et al., 2003) and in Phanerozoic ophiolites,
where they occur within the sheeted dike complexes and
the overlying lavas (Gillis and Banerjee, 2000; Gilgen
et al., 2016).
Spilites and epidosite are thus present in a variety of tec-
tonic settings and fluid flow regimes. In the in-situ oceanic
crust, spilite alteration has been traditionally ascribed to
water–rock interaction along deep downflow paths of
hydrothermal circulation cells (e.g., Mottl, 1983), but for-
mation during upflow has also been suggested (e.g.,
Coogan, 2008, Heft et al., 2008, Alt et al., 2010). Epidosite
alteration is thought to occur in the deep upflow zones of
circulation cells (Richardson et al., 1987; Schiffman and
Smith, 1988; Galley, 1993; Weber et al., 2021). Both altered
rock types have been proposed as sources of the metals in
VMS deposits (Richardson et al., 1987; Alt et al., 2010;
Jowitt et al., 2012; Patten et al., 2017).
The physicochemical properties of the fluid that causes
spilite alteration are constrained by analyses of seafloor
hydrothermal vent fluids. However, as vent fluids are typi-
cally modified by mineral precipitation and boiling at or
near the seafloor, important features of the deep fluid
remain poorly defined by direct observations, such as the
concentrations of the key elements Mg and Fe. Much less
is known about the properties of the fluid that causes epi-
dosite alteration. Opinions vary regarding its phase state
(i.e., whether a one-phase or a two-phase fluid system is
involved) as well as its salinity and density (detailed below).
Its major-element composition can be partly constrained by
thermodynamic modelling, but such models require the
basic fluid properties as input (e.g. Weber et al., 2021).
Beyond the current reach of fluid sampling at the sea-
floor, direct information on the properties of deep spilitiz-
ing and epidotizing fluids (also referred to hereafter
simply as ‘‘spilite fluidsand ‘‘epidosite fluids) can be
gained from fluid inclusion analyses, as demonstrated by
numerous studies of in-situ crust and ophiolites (reviewed
below). In the present study we verify published fluid inclu-
sion evidence from the Semail and Troodos ophiolites by
conducting new analyses from the same outcrops as those
sampled in previous studies. Further, as the main contribu-
tion of this study, we expand the available data set by pre-
senting analyses of fluid inclusions and their coeval mineral
assemblages in new spilite and epidosite samples from the
volcanic sequence of the Semail ophiolite. This allows us
to reconstruct the phase states, salinities, densities and
major element concentrations of the spilite and epidosite
fluids, as well as the depth, fluid pressure and temperature
at which the fluids were trapped as inclusions. Thus, the
characterizations of fluids and mineral assemblages that
we obtain correspond to well defined thermodynamic states
that are independent of the tectonic setting in which the
alteration occurs. Our results can therefore be used as cali-
bration targets for future numerical simulations of reactive
transport through oceanic crust in a variety of tectonic
environments, ancient and modern.
2. GEOLOGICAL SETTING
2.1. Geology of the Semail and Troodos ophiolites
The Semail oceanic crust formed in the southeastern
Tethys Ocean (Lippard et al. 1986; Searle and Cox, 1999)
in just a 1 million year interval (96.1–95.2 Ma; Rioux
et al., 2021) during the Cenomanian. Following its obduc-
tion the ophiolite was warped into a broad NW–SE-
trending antiform (Fig. 1A), tilting the crustal layering such
that erosion has now exposed oblique cross-sections in
outcrop.
The lower crust is 0.5–4 km thick and is composed dom-
inantly of layered gabbros capped by a continuous sheet of
isotropic gabbros including sparse lenses and stocks
of ‘‘plagiogranite(i.e., members of the tonalite–
trondhjemite–diorite suite; Fig. 1B). As the isotropic bodies
2 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
are the highest units in the lower crust, Lippard et al. (1986)
termed them ‘‘High Level Intrusives.Haase et al. (2016)
referred to them as the ‘‘Phase 1 (P1)generation.
The base of the upper crust is a 1–1.5 km thick sheeted
dike complex (SDC) with MORB-like major element com-
position (Fig. 2). This overlies the High Level Intrusives
and grades upwards into comagmatic basalt to basaltic–an-
desite lavas (pillows, feeder dikes and rarer sheet flows) of
the Geotimes unit (Alabaster et al., 1982). All of the rocks
mentioned above are thought to have formed at a fast
spreading axis along an oceanic ridge, which was situated
from its outset above a nascent subduction zone (e.g.,
Pearce et al., 1981; Belgrano and Diamond, 2019, and ref-
erences therein).
The axial Geotimes lavas transition upwards to primi-
tive, mainly off-axis basaltic lavas of the Lasail unit, which
are in turn overlain by up to 2 km of post-axis (Phase 2) pil-
low lavas, sheet flows, sills and feeder dikes of mostly basal-
tic to andesitic composition belonging to the Tholeiitic
Alley and Boninitic Alley units (Alabaster et al., 1982;
Belgrano et al., 2019). During Tholeiitic Alley volcanism,
stocks of gabbro and plagiogranite intruded throughout
the length of the ophiolite at all crustal levels, from the deep
layered plutonic series through to the shallow Boninitic
Alley lavas (Lippard et al., 1986; Belgrano et al., 2019).
Owing to their timing, Lippard et al. (1986) termed these
stocks ‘‘Late Intrusives, and Haase et al. (2016) referred
to them as the ‘‘Phase 2 (P2)generation. Herein we refer
to the two generations of gabbros, plagiogranites and lavas
in the ophiolite simply as ‘‘axialand ‘‘post-axial.
Volcanogenic massive sulfide (VMS) deposits in the
Semail ophiolite occur in all four volcanic units throughout
the history of late axial to late post-axial volcanism. All the
deposits are dominated by copper with only minor zinc,
regardless of the host lava unit, but the deposits within
Boninitic Alley are enriched in gold (Gilgen et al., 2014).
The oceanic crust in the Troodos ophiolite has a layered
structure and an early axial-spreading origin similar to the
Semail ophiolite (Gass, 1980). In Troodos, however, the
axial gabbros, the overlying sheeted dike complex, the
dike–lava transitional Basal Group and the overlying
Lower Lavas all have arc-tholeiitic BADR compositions
rather than MORB-like, owing to stronger influence from
an underlying subduction zone (Robinson et al., 1983).
The Upper Lavas in Troodos are boninitic (as in Semail)
to high-Mg andesitic in composition (Pearce and
Robinson, 2010). Whereas at least two major generations
and stratigraphic settings of plagiogranite are present in
Fig. 1. Geological maps of the Semail ophiolite showing sample locations of this study: (A) Geological overview (modified after Gilgen et al.,
2014); (B) North-eastern flank of the Semail ophiolite (modified after Belgrano et al., 2019) with locations of sampled plagiogranites, spilites,
epidosites and post-epidosite veins. (C) Halayn block with locations of sampled plagiogranite, epidosite and post-epidosite veins (modified
after Belgrano et al., 2019).
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 3
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Semail, plagiogranites in Troodos occur only in the gabbros
and in the sheeted dike complex but not in the lavas (e.g.,
Gass, 1980;Marien et al., 2019). Similarly to Semail,
VMS deposits occur throughout the history of volcanism
(Adamides, 2010).
2.2. Spilite alteration in the Semail and Troodos upper crust
The vast majority of sheeted dikes and lavas in the
Semail ophiolite show pervasive hydrothermal alteration,
dominantly to spilite. The SDC and the Geotimes and
Lasail lavas all display greenschist-facies alteration (see Sec-
tion 5.1 for details), with glass being extremely rare
(Kusano et al., 2017; Belgrano et al., 2019). The Tholeiitic
Alley and Boninitic Alley lavas also contain chlorite, albite
and quartz even locally up to the top of the volcanic
sequence, with smectite, prehnite and zeolites attesting to
a lower grade overprint (Alabaster and Pearce, 1985;
Pflumio, 1991; Belgrano et al., 2019). We focus the present
study on the axial SDC and Geotimes lavas that exhibit
greenschist-facies spilite alteration, including spilitized and
epidotized plagiogranites of post-axial age that sit within
these units. According to radiometric dating by Rioux
et al. (2021), the axial magmatism occurred between 96.1
and 95.6 Ma and was followed without a significant pause
by post-axial magmatism between 95.6 and 95.2 Ma. It is
thus possible to view the large-scale hydrothermal alter-
ation of the ophiolite as one extended regime through time.
Hydrothermal–metamorphic alteration in the Troodos
ophiolite (Gillis and Robinson, 1990) shows a different ver-
tical distribution compared to Semail. The lavas are limited
to chlorite–smectite grade, with chlorite–albite–quartz spi-
lites appearing in the upper SDC, and actinolite-bearing
spilites appearing towards its base. In the Troodos
literature spilitized dikes are referred to as ‘‘diabases
(e.g., Cann et al., 2015).
2.3. Epidosite alteration in the Semail and Troodos upper
crust
In the Semail ophiolite, massive epidosites are occasion-
ally present within the SDC with extents up to 0.15 km
2
.
Such epidosites were previously thought to be confined to
the base of the SDC (Nehlig et al., 1994; Juteau et al.,
2000) as in Troodos, but Gilgen et al. (2016) showed that
essentially identical epidosites occur more frequently and
with much larger outcrop extents (up to 1 km
2
) in the over-
lying lava units. In all units the epidosites overprint precur-
sor spilite alteration (Weber et al., 2021). All these
epidosites are accompanied by sparse epidote–quartz veins.
Numerous plagiogranite intrusions (both axial and post-
axial) host patchy epidosite alteration (Nehlig et al., 1994;
Gilgen et al., 2016). Nehlig (1991) suggested that each epi-
dotized dyke in the SDC was altered individually soon after
its intrusion, during active seafloor spreading. In contrast,
Gilgen et al. (2016) showed from cross-cutting relations of
dikes with known volcanostratigraphic affinity that most,
and possibly all, epidosites in the Semail ophiolite formed
after axial spreading had ceased. The epidosite alteration
overprinted the upper crustal sequence and its enclosed
Fig. 2. Upper crustal stratigraphy of the Semail ophiolite based on Belgrano and Diamond (2019). SDC – Sheeted Dike Complex; Geotimes
(axial), Lasail (off-axial), Tholeiitic Alley (post-axial) and Boninitic Alley (post-axial) – basaltic to basaltic-andesite lava units, mostly altered
to spilite. Spilite and epidosite sample locations for this study: 1 – Highway 7, 2 – Bani Umar North, 3 – Rusays, 4 – Wadi Rajmi, 5 – Aarja-
Bayda, 6 – Bani Umar North, 7 – Wadi Hawqayn, 8 – Jiltah, 9 – Wadi Hawqayn, 10 – Al Malah, 11 – Ajeeb, 12 Rajmi, 13 – Rusays.
Numerous other known epidosites in the SDC, Geotimes and Lasail lavas (Gilgen et al., 2016) are not shown.
4 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
post-axis intrusions once the Lasail unit and some of the
overlying Tholeiitic Alley and Boninitic Alley units had
already erupted (Fig. 2).
In the Troodos ophiolite, massive epidosites occur abun-
dantly within an area of 10 km
2
as large, pervasive
replacements of the SDC and the upper gabbros, and as
patchy alteration of plagiogranites at the gabbro–SDC
boundary (Richardson et al., 1987; Schiffman and Smith,
1988). Within these areas, the percentage of end-member
epidosites is small (e.g., Jowitt et al., 2012). The alteration
in the plagiogranites affects large volumes of rock but indi-
vidual epidosite bodies within them are often only pods of a
few dm
3
to cm
3
in volume. In contrast to Semail, epidosites
in the Troodos volcanic sequence are extremely rare,
despite ample outcrops. A paleomagnetic study by Varga
et al. (1999) suggested the epidosites formed prior to cessa-
tion of the seafloor spreading that created the SDC. Based
on cross-cutting relations of dikes with striped epidosite
alteration, Cann et al. (2015) argued that each individual
dike had been epidotized immediately upon its intrusion,
before the next adjacent dike intruded the SDC, as sug-
gested by Nehlig (1991) for Semail. However, Gilgen
et al. (2016, p. 207) pointed out that the same cross-
cutting relations in Troodos can be alternatively explained
by a single influx of epidotizing fluid after large segments
of the heterogeneously permeable SDC had already formed.
3. PREVIOUS FLUID INCLUSION STUDIES OF
OCEANIC SPILITES AND EPIDOSITES
3.1. Fluid inclusions in spilites
Previous fluid inclusion studies agree on the basic prop-
erties of the spilite fluid that has altered the SDC and SDC-
lava transition in in-situ oceanic crust (Fig. 3A, field 1; e.g.,
Gallinatti, 1984; Delaney et al., 1987; Saccocia and Gillis,
1995; Heft et al., 2008; Alt et al., 2010; Castelain et al.,
2014) and in the Troodos ophiolite (Kelley et al., 1992;
Gillis, 2002). Quartz is the only host mineral that has
yielded tractable inclusions, but even these are typically
very small. Owing to petrographic limitations, few studies
have been able to identify primary inclusions from the
abundant secondary and other inclusions of equivocal age
with respect to quartz. Only aqueous liquid-dominated liq-
uid–vapor (LV) inclusions have been found, suggesting that
only one liquid phase was present in the rocks during spilite
alteration and fluid inclusion entrapment. The salinities of
the inclusions vary from 0.1 to 10.1 wt.% NaCl
eq
, straddling
the salinity of modern seawater (3.0–3.8 NaCl
eq
, or 3.1–3.9
TDS; Levitus et al., 1994). This variation is attributed to
high temperature liquid–vapor separation of seawater and
partial remixing at depth, prior to entrapment of the inclu-
sions (discussed further in Section 6.3.2). Homogenization
temperatures (T
h
(LV?L)) vary over a broad, 320 degree
range from 130 to 450 °C, implying that fluid was trapped
over a similarly broad range of temperatures. Temperatures
of entrapment (T
trap
) are typically 10–40 °C higher than the
measured T
h
values, as estimated by adding a correction to
T
h
based on the hydrostatic pressure of cold seawater at the
known or estimated depth of entrapment. Presumably the
spilite alteration assemblage, involving chlorite, quartz,
albite and often actinolite, formed in the upper half of the
T
trap
range.To our knowledge there is so far no record of
hypersaline (halite-bearing) brines in the SDC or the over-
lying lavas from the in-situ or ophiolite crust. However,
liquid–vapor–halite (LVH) inclusions with salinities of
31–61 wt.% NaCl
eq
and conjugate vapor-dominated LV
inclusions with very low salinities (1–4 wt.% NaCl
eq
) are
abundant in plagiogranites, felsic dikes and some gabbros
(Fig. 3A, field 4: Kelley et al., 1992;Vanko et al., 1988,
Vanko et al., 1992;Kelley and Fru
¨h-Green, 2001; Alt
et al., 2010; Castelain et al., 2014; Bali et al., 2020). Both
primary and secondary inclusions occur in igneous quartz
and apatite, and in quartz ± epidote veins. The common
occurrence of hypersaline brine and vapor in felsic intru-
sives and not in nearby gabbros suggests that they are
exsolved magmatic fluids rather than seawater that has
undergone liquid–vapor separation (Kelley and Fru
¨h-
Green, 2001). The plutonic rocks also contain abundant
primary and secondary LV inclusions of the same spilite
type as in the SDC and lavas (Vanko et al., 1988,Vanko
et al., 1992;Nehlig, 1989, 1991; Alt et al., 2010; Castelain
et al., 2014), with salinities of 0.5–9.2 wt.% NaCl
eq
and
homogenization temperatures of 125 °C to over 500 °C
(Fig. 3A, field 2). Less common are LV inclusions with
salinities of 2–23 wt.% NaCl, which almost bridge the
low-salinity LV and hypersaline fluid types (Fig. 3A, field
3; Kelley et al., 1992;Vanko et al., 1992;Castelain et al.,
2014). These may be due to condensation and remixing of
the magmatic fluids (Kelley and Fru
¨h-Green, 2001).
3.2. Fluid inclusions in epidosites
Previous fluid inclusion studies on epidosites are
restricted almost exclusively to the Troodos and Semail
ophiolites (Fig. 3B). Two contrasting types of fluids have
been found associated with the epidosites and deduced to
have caused the alteration. The first has microthermometric
properties (Fig. 3B, field 1) that are very similar to the spi-
lite fluids reviewed above (Fig. 3A, field 1). Quartz within
the massive epidosites contains primary and pseudosec-
ondary aqueous LV inclusions with salinities of 3–7 wt.%
NaCl
eq
and T
h
(LV?L) between 200 and 410 °C
(Richardson et al., 1987; Schiffman and Smith, 1988).
Essentially identical inclusions are present in epidote–
quartz veins that cross-cut nodular massive epidosites in
the Semail SDC (1.6–5.7 wt.% NaCl
eq
;T
h
= 258–396 °C;
Nehlig and Juteau, 1988) and in a breccia of quartz-
cemented epidotized clasts from the Oceanographer Trans-
form fault, Mid-Atlantic Ridge (3.2–8.0 wt.% NaCl
eq
;
Vanko et al., 1992). Identical secondary inclusions occur
in epidotized plagiogranites in Troodos, but their timing
with respect to the epidotization is unclear (Kelley and
Robinson, 1990).
The second type of inclusions associated with epidosites
are hypersaline (LVH) brines and coexisting, low salinity
vapor-dominated LV inclusions. Cowan and Cann (1988)
reported such inclusions in quartz and epidote in an epido-
tized plagiogranite in the Troodos SDC, and interpreted
them as the product of liquid–vapor separation of seawater.
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 5
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Fig. 3. Fluid inclusion salinities versus homogenization temperatures (T
h
) from the literature and this study: (A) Spilite fluids in lavas, sheeted
dikes and plutonics: 1)Gallinatti, 1984; Delaney et al., 1987; Kelley et al., 1992; Saccocia and Gillis, 1995; Gillis, 2002, Heft et al., 2008; Alt
et al., 2010; Castelain et al., 2014.2)Nehlig, 1989, 1991;Vanko et al., 1988,Vanko et al., 1992;Kelley and Fru
¨h-Green, 2001; Alt et al., 2010;
Castelain et al., 2014.3)Kelley et al., 1992;Vanko et al., 1992;Kelley and Fru
¨h-Green, 2001; Castelain et al., 2014.4) Hypersaline brines, not
necessarily associated with spilite alteration, are shown for reference (Kelley et al., 1992;Vanko et al., 1988,Vanko et al., 1992;Alt et al., 2010;
Castelain et al., 2014; Bali et al., 2020). (B) Epidosite fluid and later vein fluids: 1)Richardson et al., 1987; Schiffman and Smith, 1988;
Bettison-Varga et al., 1995; Nehlig et al., 1994; Nehlig and Juteau, 1988.2)Kelley et al., 1992; Cowan and Cann, 1988.3, 4) Kelley et al.,
1992; Cowan and Cann, 1988; Richardson et al., 1987; Juteau et al., 2000. Abbreviations: LV – liquid–vapor; LVH – liquid–vapor–halite;
Ep – epidote; Qz – quartz; SDC – sheeted dike complex.
6 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Detailed petrographic descriptions were not provided.
Identical fluid inclusions were found by Kelley and
Robinson (1990) and Kelley et al. (1992) in small epidosite
pods in the Troodos plagiogranites and were contrastingly
interpreted to represent magmatic fluids exsolved from
the plagiogranite magma. The same high salinity inclusions
occur in epidote–quartz veins in epidotized axial pla-
giogranites in the Semail SDC (Juteau et al., 2000).
Juteau et al. (2000) attributed the origin of the hypersaline
brines to unmixing of seawater or to magmatic exsolution,
and suggested that the fluids fed the Semail VMS deposits
during axial spreading. These findings and interpretations
are discussed in light of our new evidence in Section 6.4.
4. FLUID INCLUSION METHODOLOGY, SAMPLING
AND ANALYTICAL METHODS
Compared to the previous studies reviewed above, the
present work employs a modified methodology and addi-
tional analytical methods, as follows.
4.1. Fluid inclusion methodology
Our approach to identify inclusions that trapped the
spilitizing and epidotizing fluids involves the following four
steps. First, we have confined our analyses to groups of
fluid inclusions that can be proven to be coeval based on
direct petrographic evidence at the thin-section scale. Such
demonstrably coeval groups are termed ‘‘fluid inclusion
assemblages(Diamond, 1990). Comparing the properties
of several inclusions in an individual assemblage provides
the basis to recognize and disregard any inclusions modified
by post-entrapment processes (e.g., necking down,
Roedder, 1981) and to identify the phase state of the fluid
during trapping. Thus, assemblages of unmodified inclu-
sions that display uniform phase-volume proportions at
laboratory temperature are interpreted to have been
trapped from a pore fluid that consisted of a single fluid phase
(‘‘homogeneous entrapment). In contrast, assemblages of
inclusions that display variable phase-volume proportions
are interpreted to have been trapped from a pore fluid that
contained two or more immiscible phases (‘‘heterogeneous
entrapment; see Diamond, 2003 for principles).
Second, we have used the petrographic criteria in Roedder
(1984) to established the relative ages of fluid inclusion
assemblages with respect to host-mineral growth: primary
in crystal growth zones, secondary in post-growth healed
fractures, and pseudosecondary in syn-growth healed frac-
tures. As any fracture through a crystal can terminate by
chance at an internal growth horizon when viewed in a petro-
graphic section, we have taken care to classify healed frac-
tures as pseudosecondary only when several fractures can
be seen terminating at the same growth horizon.
Third, to make mineral growth zonation visible and
thereby permit recognition of primary inclusion assem-
blages, we have applied a variant of SEM-
cathodoluminescence (CL) petrography using a secondary
electron detector under variable pressure (‘‘VPSE imaging;
Lambrecht and Diamond, 2014). This technique reveals
growth zonation in quartz caused by variations in contents
of trace elements and other defects that generate lumines-
cence under the electron beam of the SEM (e.g., Go
¨tte
et al., 2011).
Fourth, we recognize that the studied ophiolites have
undergone protracted histories of water–rock interaction
and brittle deformation during sub-seafloor hydrothermal
alteration and subsequent obduction and exhumation, as
evidenced by later generations of hydrothermal veins that
cross-cut the spilites and epidosites. To ensure that we
can distinguish later fluid inclusions from those that were
trapped during the alteration events of interest, we have
established in the field the sequence of vein generations in
the upper crust of the Semail ophiolite and we have ana-
lyzed primary or pseudosecondary fluid inclusions in each
vein generation.
4.2. Sampling
Locations of the Semail samples analyzed in this study
are shown in Fig. 1. Their volcanostratigraphic and outcrop
settings are shown in Figs. 2 and 4, and details are listed in
Tables 1 and Supplementary Material 1. To characterize
spilite fluids in the Semail ophiolite we sampled
quartz ± chlorite-bearing amygdales in spilitized Geotimes
lavas, and quartz-bearing miarolitic cavities in spilitized
axial and post-axial plagiogranites. For the epidosite fluids
we sampled massive epidosite and associated epidote–
quartz veins within the SDC (including the localities sam-
pled by Nehlig, 1994, and Nehlig and Juteau, 1988), within
the Geotimes pillow lavas (at localities described by Gilgen
et al., 2016), and within locally epidotized axial and post-
axial plagiogranites (including localities sampled by
Juteau et al., 2000). The sampled massive epidosites are
of the rarer end-member type, consisting of epidote, quartz,
titanite and Fe-oxides, with virtually no relict Mg-bearing
minerals inherited from the precursor spilites. Although
most of the sampled rocks are of axial magmatic age, it is
important to realize that some of the sampled spilite alter-
ation and most of the sampled epidosite alteration occurred
during post-axial magmatism.
In addition to the spilites and epidosites, we sampled all
the known generations of cross-cutting (post-epidosite)
hydrothermal veins: prehnite–quartz veins, quartz–
hematite–pyrite (Q’) veins, and calcite veins (Table 1;
details in Supplementary Material 1).
For comparison with the Troodos ophiolite we visited
the sites of earlier studies and obtained epidosites from
SDC locality 5 of Richardson et al. (1987; our sample
Co51) and from epidotized miarolitic cavities in the altered
plagiogranite 0.5–1 km east of Platanistasa sampled by
Cowan and Cann (1988) and by Kelley et al. (1992; their
locality 9; our sample LR18-Tr06), as detailed in Supple-
mentary Material 1.
4.3. Analytical methods
4.3.1. Optical petrography
Thin sections and doubly-polished thick sections were
examined petrographically using an Olympus BX51 polar-
izing microscope equipped with a range of objective lenses
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 7
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
up to 100x, and an Olympus UC90 digital camera. In order
to interpret the phase state at fluid trapping, the variation in
volume fractions of vapor (denoted u
vap
) and of captured
solids (u
solid
) within each assemblage was recorded by esti-
mating the approximate area fraction of the phases. These
interpretations were later checked against the spread in
homogenization temperatures of each assemblage, which
directly and accurately reflects the spread in u
vap
values
(Bakker and Diamond, 2006;Diamond, 2003).
4.3.2. Cathodoluminescence (SEM-VPSE) imaging
Uncoated, polished thin- and thick sections were used
for CL imaging following the SEM-VPSE method of
Lambrecht and Diamond (2014). A Zeiss EVO 50 Environ-
mental Scanning Electron Microscope (SEM) was used to
acquire grey-scale images of the CL intensity generated by
a 14 kV, 2nA electron beam with a spot size of 500 nm
diameter, under 12 Pa air pressure.
4.3.3. Fluid inclusion microthermometry
Microthermometry was performed on doubly polished
rock sections of 80–200 mm thickness using a Linkam
THSG 600 heating–freezing stage mounted on an Olympus
BX 51 microscope equipped with an Olympus 100x
LMPlanFL objective lens of 0.95 NA and a matching sub-
stage condenser. The stage was calibrated against phase
Fig. 4. Schematic vertical stratigraphy in the Semail ophiolite from the lower crustal gabbros up through the sheeted dike complex to the
overlying lavas, showing the outcrop settings and types of samples used in this study. Numbers 1 to 14 refer to sample type numbers in
Table 1. Abbreviations: Ep – epidote; Qz – quartz; Act – actinolite; Py – pyrite; Tnt – titanite; Hem – hematite; Ab – albite; Chl – chlorite.
8 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Table 1
Name, type and characteristics of samples from the Semail (Oman) and Troodos (Cyprus) ophiolites analyzed in this study.
Sample
type
Timing Fluid generation Sample/ rock type Stratigraphic level Location*Sample number Mineral assemblage
1 Pre-epidosite Magmatic–
hydrothermal
Axial plagiogranites Isotropic gabbro Wadi Rajmi LR-WR-T01 Amp+Qz+Pl
Axial plagiogranite SDC Troodos LR18-Tr06 Amp+Qz+Pl
Axial plagiogranite SDC Troodos LR18-Tr15 Amp+Qz+Pl
2 Pre-epidosite Magmatic–
hydrothermal
Post-axial plagiogranites SDC Wadi Hawqayn LR18-WH-T02 Amp+Qz+Pl+Ilm
Post-axial plagiogranites SDC Jiltah LR17-J-T05 Amp+Qz+Pl+Ilm
3 Pre-epidosite Magmatic–
hydrothermal
Post-axial plagiogranites Lavas Wadi Bani Umar
North
LR17-BUN-Ta Amp+Qz+Pl+Ilm
Post-axial plagiogranites Lavas Aarja-Bayda LR17-AB-T07 Amp+Qz+Pl+Ilm
Post-axial plagiogranites Lavas Lasail South LR-TT-06 Amp+Qz+Pl+Ilm
4 Pre-epidosite Spilite Dike SDC Wadi Hawqayn AB17-32c Act+Chl+Qz
5 Pre-epidosite Spilite Amygdale in pillow rim Lavas Highway 7, km 15 LR17-BU-Sp02 Qz+Chl+Pmp+Ep in
amygdales
Amygdale in pillow rim Lavas near Highway 7, km
15
AB16-4233 Qz+Chl+Pmp+Ep in
amygdales
5 Pre-epidosite Spilite Lava drainage cavities Lavas Rusays LR17-F-Ep03 Qz
6 Pre-epidosite Spilite Axial plagiogranite Isotropic gabbro Wadi Rajmi LR-WR-T01 Qz+Act
7 Pre-epidosite Spilite Post-axial plagiogranite Lavas Wadi Bani Umar
North
LR17-BUN-Ta Qz+Act+Chl+Ep
8 Syn-
epidosite
Epidosite Massive epidosite SDC Wadi Hawqayn LR17-WH-SD-
Ep03
Ep+Act+Qz
Massive epidosite SDC Troodos Co-51 Ep+Chl+Qz
9 Syn-
epidosite
Epidosite Lava drainage cavity in massive
epidosite
Lavas Ajeeb LR-A-Ep02 Ep+Qz+Tnt+Hem
Lava drainage cavity in massive
epidosite
Lavas Ajeeb LR-A-Ep15 Ep+Qz+Tnt+Hem
Lava drainage cavity in massive
epidosite
Lavas Ajeeb SG13-20-2 Ep+Qz+Tnt+Hem
Lava drainage cavity in massive
epidosite
Lavas Ajeeb LR17-A-Ep02 Ep+Qz+Tnt+Hem
Lava drainage cavity in massive
epidosite
Lavas Ajeeb LR17-A-Ep03 Ep+Qz+Tnt+Hem
Lava drainage cavity in massive
epidosite
Lavas Rusays LR17-F-Ep03 Ep+Qz+Tnt+Hem
10 Syn-
epidosite
Epidosite Ep–Qz vein SDC Wadi Hawqayn LR17-WH-SD-
Ep03
Ep+Qz+Hem
11 Syn-
epidosite
Epidosite Ep–Qz vein Lavas Ajeeb SG13-41 Ep+Qz+Tnt+Hem
12 Syn-
epidosite
Epidosite Miarolitic cavities in post-axial
plagiogranite
SDC Wadi Hawqayn LR18-WH-T02 Ep+Qz+Hem
Massive epidosite SDC Troodos LD-CY-8 Ep+Chl+Qz
Miarolitic cavities in post-axial
plagiogranite
SDC Troodos LR18-Tr06 Ep+Qz
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 9
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
transitions in synthetic, quartz-hosted H
2
O and CO
2
–H
2
O
inclusions of known composition and density (melting of
solid CO
2
at -56.6 °C, melting of H
2
O-ice at 0.0 °C and crit-
ical homogenization of H
2
O at 374.1 °C). Measurements
below room temperature are thus accurate to ±0.1 °C
and those above room temperature are accurate to ±1.5 °C.
All the analyzed inclusions contain aqueous liquid (L)
and aqueous vapor (V) phases at laboratory temperature
(T
lab
), and some contain solids (e.g., halite, H). Halite
was identified by its reaction with H
2
O to form hydrohalite
at low temperature, by the absence of a meniscus where it
touches the ordinary optical axis of quartz, and by its
absence of a Raman spectrum. After first cooling the inclu-
sions to -190 °C, two equilibrium phase transitions were
measured upon gradual heating at 0.5 °C/min: (1) the tem-
perature at which the last crystal of ice melts in the presence
of liquid and vapor, denoted T
m
(Ice); (2) the temperature at
which the inclusion phases homogenize. In LV inclusions
dominated by liquid at T
lab
this corresponds to a bubble-
point transition via shrinkage and disappearance of the
vapor bubble, denoted T
h
(LV?L). In LV inclusions dom-
inated by vapor at T
lab
, homogenization occurs via expan-
sion of the vapor bubble in a dew-point transition, denoted
T
h
(LV?V). Usually only a minimum constraint can be
placed on the dew-point temperature, owing to disappear-
ance of the L–V meniscus into the dark inclusion walls as
the true homogenization temperature is approached. All
inclusions were checked for the presence of gas-clathrates
at temperatures above T
m
(Ice).
The salinity of the inclusions based on microthermome-
try alone is expressed in NaCl equivalent concentration (wt.
% NaCl
eq
) regardless of the true solute compositions. The
salinity of LV inclusions was calculated from T
m
(Ice) and
the salinity of LVH inclusions was calculated from
T
m
(H), both using the AqSo_NaCl P–T–V
m
xcorrelation
for the binary H
2
O–NaCl system in Bakker (2018; based
on Driesner and Heinrich, 2007; Driesner, 2007).
4.3.4. Laser Raman analyses of fluid inclusions
Laser Raman analyses were conducted on a confocal
Jobin Yvon Horiba microprobe (LABRAM HR-800)
equipped with an Olympus BX41 microscope. A
532.12 nm laser was focused through an Olympus 100x
UMPlan FL objective to obtain a beam spot on the sample
of 20 mW power and 1mm diameter.
Vapor bubbles in fluid inclusions were scanned for gas
species, including H
2
O, CH
4
,CO
2
,H
2
and H
2
S, at a spec-
tral resolution of 2 cm
-1
with 20 s acquisition time. The
identity of the final solid phase to melt upon warming the
inclusions from sub-zero temperatures was determined by
mounting the heating-freezing stage on the Raman micro-
probe and scanning for the characteristic spectra of ice,
hydrohalite (NaCl2H
2
O) and solid hydrates of CaCl
2
.
Other accidentally trapped solids in the inclusions were
identified from their characteristic spectra at room
temperature.
4.3.5. Laser-ablation ICP-MS analyses of fluid inclusions
Mass ratios of elements in individual fluid inclusions in
quartz were determined by laser-ablation inductively-
13 Syn-
epidosite
Epidosite Miarolitic cavities in post-axial
plagiogranite
Lavas Wadi Bani Umar
North
LR17-BUN-Ta Ep+Qz
14 Syn-
epidosite
Distal epidosite Lava drainage cavity Lavas Al Malah LR-AM-Ep06 Ep
15 Post-
epidosite
Prehnite–quartz Vein Lavas Al Malah LR-AH-03 Prh+Qz
Vein Axial gabbros and
SDC
Wadi Haymiliyah LR18-WHim-G Prh
Lava drainage cavity Lavas Ajeeb (Hatta
Extended)
LR18-HE-Ep01 Prh+Qz
Vein Lavas Ajeeb 220122-7 Prh+Qz
Vein Lavas Yanqul SWOM-17-29 Prh+Qz
16 Post-
epidosite
Hematite–
quartz ± pyrite
Vein SDC Wadi Haymiliyah LR18-WH-SD-
Ep02
Hem+Qz+Chl+Py+Cpy
Vein Lavas Haduf 260112-11A Hem+Qz+Chl
Vein Lavas Ajeeb LR17-A-Qz01a Hem+Qz+Chl
17 Post-
epidosite
Calcite–quartz Vein SDC Wadi Hawqayn LR17-WH-SD-
Ep03
Cal+Qz+Hem
Vein Lavas Ajeeb LR17-A-Cal02a Cal+Qz+Hem
*
see Supplementary Material 1 for GPS coordinates; Mineral abbreviations: Act – actinolite; Amp – amphibole; Cal – calcite; Chl – chlorite; Cpy – chalcopyrite; Ep – epidote; Hem – hematite;
Ilm – ilmenite; Pl – plagioclase; Prh – prehnite; Pmp – pumpellyite; Py – pyrite; Qz – quartz; Tnt – titanite.
10 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Table 2
Mineral–fluid parageneses in seven fluid generations associated with miarolites in axial and post-axial plagiogranites, with spilite and epidosite alteration and with hydrothermal veining in the
Semail ophiolite.
Time?
Fluid generation
Mineral Plagiogranite melt Magmatic–hydrothermal Spilite Epidosite Prh–Qz
veins
Qz–Hem ± Py
veins
Cal–Qz
veins
Salinity T
h
[wt.% NaCl
eq
][°C]
Plagioclase –––
Albite ––– –––
K-feldspar – Rare –
Amphibole – Melt incl. –
Ilmenite –––
Apatite ––– ––– –––
White Qz* – Melt incl. –
Med.-grey Qz* – Brine ± vapor FI – Brine: 30 380–410
Dark Qz* – LV FI – 2.4–4.0 128–390
Dark Qz* – LV FI – 2.6–3.9 225–370
Twinned Qz* – LV FI – 3.4–3.7 (220)
Dark Qz* – LV FI – 2.6–2.9 100–145
Dark Qz* – LV FI – 0.5–1.9 62–81
Epidote ––– – LV FI –
Titanite – LV FI – –––
Rutile –––
Magnetite ––– –––
Hematite ––– ––– ––– –––
Actinolite – LV FI –
Chlorite –––
Prehnite –––
Calcite
Pyrite –––
Seawater 3.1–3.9
* Degree of brightness in greyscale CL image; Abbreviations: Qz – quartz; Rt – rutile; Hem – hematite; Prh – prehnite; LV – liquid-vapor; FI – fluid inclusions;
for details see Supplementary
Material 1.
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 11
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
coupled-plasma mass-spectrometry (LA-ICP-MS). Fluid
inclusions in epidote were not analyzed owing to the high
background of Ca in the host. A Lambda Physik
GeoLas-Pro 193 nm ArF excimer laser system was used
in combination with a Perkin Elmer ELAN DRC-e or an
Agilent 7900 quadrupole mass spectrometer following the
procedure in Pettke et al. (2012). Analyses were made for
7
Li,
11
B,
23
Na,
25
Mg,
29
Si,
35
Cl,
39
K,
43
Ca,
57
Fe,
79
Br,
85
Rb,
88
Sr and
137
Ba, as calibrated against the standards
GSD-1G (Jochum et al., 2011) and Sca-17 (Seo et al.,
2011; Fusswinkel et al., 2018).
Ablation profiles were processed using the SILLS soft-
ware (Guillong et al., 2008;Figs. S10, S11), yielding mass
ratios or limits of detection (LOD) where signals are not
significant. The LOD values pertain to individual fluid
inclusions depending on inclusion size, depth, ablation
behaviour, etc. in addition to instrumental settings. For
a given element, the lowest LOD among the inclusions
analyzed in an assemblage is accordingly the most rele-
vant. Nevertheless, if the inclusion with the lowest LOD
has unfavorable fluid inclusion properties, its calculated
LOD may exceed the element concentration derived from
other inclusions with statistically significant element sig-
nals. Occasional outliers were filtered from the final data
set by applying the conventional Extreme Studentized
Deviation test, which assumes Gaussian distribution. To
avoid skewing, median values are used where eight or less
inclusions were accepted per fluid inclusion assemblage.
Since Na is the dominant cation in all the analyzed inclu-
sions, and since the element concentrations in the inclu-
sions can be later calculated using the known NaCl
eq
salinity determined by microthermometry, the relative
concentrations of the analytes are expressed as mass ratios
with respect to Na.
4.3.6. Calculation of bulk fluid inclusion compositions
For the samples that yielded useful element ratios by
LA-ICP-MS, bulk fluid compositions were calculated as
follows. Since Cl, Na and Ca dominate all inclusions with
salinities 5.3 wt.% NaCl
eq
(Section 5.3), and since the pre-
cision of the Cl analyses is too limited to constrain charge-
balance, the analyzed cations were artificially balanced with
chloride and then grouped into 1:1 versus 1:2 chloride salts.
Owing to the low salinities of these inclusions and to the
dominance of Na and Ca, it can be assumed that the 1:1
salts have the same colligative properties as NaCl and that
the 1:2 salts have the same properties as CaCl
2
(Figs. S12
and S13). The concentrations of the solutes were then calcu-
lated from the ratio of salt types plotted on the relevant
T
m
(Ice) isotherm in the H
2
O–NaCl–CaCl
2
model system,
using the liquidus diagram of Steele-Maclnnis et al.
(2011). This calculation also permitted conversion of
NaCl
eq
into salinity in terms of total dissolved solids (TDS).
In the hypersaline (LVH) inclusions the cations were
similarly balanced with Cl and grouped into 1:1 and 1:2
salts. The dominant solutes are Cl, Na and K but there is
also significant Fe. To account for the colligative effects
of Fe on the 1:2 group, the mass ratio of salt groups was
plotted on the relevant T
m
(H) isotherm of the H
2
O–
NaCl–FeCl
2
model halite liquidus (Lecumberri-Sanchez
et al., 2015), thereby permitting calculation of bulk solute
concentrations (Fig. S14).
5. RESULTS
5.1. Petrography of altered upper crustal rocks
The following petrographic descriptions provide the
mineralogical and temporal context for the fluid inclusion
Table 3
Mineral–fluid parageneses in four fluid generations associated with plagiogranite miarolites and with spilite and epidosite alteration in the
Troodos ophiolite.
– Time ?
Fluid generation
Mineral Plagiogranite melt Magmatic–hydrothermal Spilite Epidosite Fluid salinity
[wt.% NaCl
eq
]
T
h
[°C]
Plagioclase –––
Albite –––
K-feldspar – Rare –
Amphibole –––
Ilmenite –––
Apatite ––– ––– –––
White Qz* – Melt incl. –
Med. grey Qz* – Brine+vapor FI – Brine: >30 340–356
Dark Qz* – LV FI – 5.7 318–323
Dark Qz* – LV FI – 3.2 272–283
Epidote – LV FI – 3.1–3.2 275–280
Titanite
Magnetite
Hematite ––– –––
Actinolite –––
Chlorite –––
Seawater 3.1–3.9
* Degree of brightness in greyscale Cl image; Abbreviations: Qz – quartz; LV – liquid-vapor; FI – fluid inclusions.
12 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
analyses presented in Section 5.2. Emphasis is placed on
identifying the temporal sequence of hydrothermal mineral
assemblages in each of the sampled rock units. Tables 2 and
3summarize the paragenetic sequences comprising the
alteration types observed in the Semail and Troodos ophi-
olites, revealing the mineral assemblages coeval with the
analyzed fluid inclusions.
5.1.1. Sheeted Dike Complex (SDC)
Most dikes in the SDC are spilitized diabases (Nehlig
et al., 1994; Miyashita et al., 2003) displaying alteration
assemblages involving albite + chlorite with varying
amounts of hydrothermal actinolite and minor quartz, epi-
dote, titanite and magnetite. Relict igneous augite and
titanomagnetite are often preserved. This assemblage is
locally completely replaced by massive, granoblastic
epidosites (Fig. 5A) comprising intergrowths of euhedral
epidote and quartz with accessory titanite and hematite or
occasionally magnetite, often containing relicts of spilitic
actinolite (Nehlig, 1989; Gilgen et al., 2016). Associated epi-
dote–quartz veins up to 2 cm wide cross-cut the massive
epidosite bodies and the adjacent spilites (Fig. 5B). The
veins are mostly compact but locally vuggy, consisting of
coevally grown euhedral epidote and quartz. These miner-
als are often partially overgrown by later, genetically unre-
lated calcite and quartz. This late quartz is readily
distinguished from the epidote-generation quartz by petro-
graphic criteria.
5.1.2. Lavas
Spilitized Geotimes pillow lavas consist of hydrothermal
chlorite + albite + titanite + hematite ± quartz ± epidote ±
Fig. 5. Field and microphotographs of massive epidosites in dikes of the SDC and associated cross-cutting Ep–Qz veins (Sample: LR17-WH-
SD-Ep03) from the Semail ophiolite. (A) Completely epidotized dike core. (B) Epidotized dikes hosting Ep–Qz veins cross-cut by later Qz-
Hem and calcite veins. (C) Close-up of an Ep–Qz vein with euhedral Ep + Qz. (D) Close-up of (C): Secondary, homogeneously trapped LV
fluid inclusions in quartz. Abbreviations: Ep – epidote; Qz – quartz; Hem – hematite; SDC – sheeted dike complex.
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 13
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
submicroscopic needles of actinolite. Relict igneous augite
and titanomagnetite are common but relict plagioclase is
only occasionally preserved (Alabaster and Pearce, 1985;
Gilgen et al., 2016). Vesicles (mm–cm diameter) are often
empty but in certain areas vesicles and lava drainage cavi-
ties in the pillows are lined by mm-size euhedral quartz
crystals intergrown with chlorite, indicating that they
belong to the spilite assemblage (Fig. 6A, B). In spilitized
lavas overprinted by weak or intense epidosite, the precur-
sor cavity quartz is often overgrown by euhedral epidote
and a second generation of quartz belonging to the epi-
dosite assemblage (Fig. 7A). Where the two generations
are present in the same thin section they can be distin-
guished systematically not only by their mineral inter-
growths but also by the slightly brighter CL intensity of
the precursor spilite-generation quartz (Fig. 7E). The CL
images show that crystal faces of the spilite-generation
quartz occasionally underwent partial dissolution prior to
being overgrown by epidosite-generation quartz.
Sparse epidote–quartz veins up to 2 cm wide, with the
same features as those in the sheeted dikes, cross-cut the
epidotized pillows and often reach out into the surrounding
spilitized lavas. Post-epidosite fillings of vesicles and drai-
nage cavities include pumpellyite, prehnite and zeolite.
5.1.3. Plagiogranites
The paragenetic sequence of igneous and hydrothermal
minerals observed in the Semail plagiogranites is shown
in the first two columns of Table 2 (based on petrographic
relationships in Figs. S1–S4). The igneous mineral assem-
blages of the axial and post-axial plagiogranites are very
similar. Both consist of plagioclase, edenitic amphibole
and minor quartz with rare K-feldspar and clinopyroxene
plus accessory ilmenite and apatite. Some post-axial pla-
giogranites also contain magnetite. The interstitial matrix
quartz in both generations exhibits bright CL (light-grey
to white) typical of igneous quartz in granitoids worldwide
(e.g., Landtwing and Pettke, 2005). The quartz is also often
intergrown granophyrically with plagioclase. Both quartz
and amphibole contain former melt inclusions (now crystal-
lized), as described in Section 5.2.1.
Huge numbers of regularly spaced miarolitic cavities
are present in some areas of the plagiogranite stocks.
The cavities have equant to lensoid shapes and are up
to several cm in diameter. Some elongate cavities are
linked to adjacent cavities, but the three-dimensional out-
crop conditions are clear enough to conclude that the
majority of cavities are unconnected. The cavities are bor-
dered by pegmatitic intergrowths of coarse igneous plagio-
clase and quartz, containing almost none of the edenitic
amphibole that otherwise makes up about 20 vol.% of
the plagiogranite matrix. The cavities are partly filled by
freely grown hydrothermal actinolite and euhedral quartz.
The cores of the quartz crystals contain apatite and melt
inclusions and exhibit intense CL, as does igneous quartz
in the rock matrix.
Fig. 6. (A, B) Field photographs of spilite-generation quartz in amygdales in Geotimes lavas. (C) Close-up of B: Microphotograph of
amygdale quartz with growth zonation. (D) Primary LV fluid inclusions along quartz growth zone. Abbreviations: Qz – quartz; LV – liquid-
vapor.
14 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Both the igneous quartz crystals in the matrix and in the
miarolitic cavities are epitaxially overgrown by quartz with
medium-grey CL that we refer to as ‘‘magmatic–
hydrothermal. The igneous quartz shows CL evidence
for partial dissolution prior to being overgrown by the
magmatic–hydrothermal quartz. Elongate apatite is the
Fig. 7. Field and microphotographs of massive epidosites in axial Geotimes lavas and associated Ep–Qz veins in the Semail ophiolite: (A)
Lava drainage cavities filled with euhedral Ep + Qz (Sample: LR17-A-Ep03). (B) Ep–Qz vein in massive epidosite overprinted by a later Hem–
Qz vein assemblage (Sample: SG13-41). (C) Euhedral Ep crystals in Ep–Qz veins containing (D) primary LV fluid inclusions. (E) Transmitted
light and SEM-VPSE image of euhedral Qz + Ep and (F) pseudosecondary LV inclusions with uniform vapor fractions. Abbreviations as in
Figs. 5 and 6.
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 15
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
only other mineral we have found coeval with the mag-
matic–hydrothermal quartz.
All the plagiogranites are to some degree pervasively
spilitized, as manifested by alteration of plagioclase to
albite, of clinopyroxene and igneous amphibole to
hydrothermal actinolite and chlorite, and of magnetite
and ilmenite to titanite, hematite and rarely rutile.
Hydrothermal quartz is also part of the spilite assemblage,
most notably overgrowing magmatic–hydrothermal quartz
epitaxially in the miarolitic cavities. This spilite-generation
quartz is dark in CL images, allowing the three generations
of quartz to be distinguished. The CL of the spilite-
generation quartz reveals that it also occurs pervasively in
the rock matrix as coatings or fillings of microfractures
within the earlier quartz generations. Magmatic–hydrother-
mal quartz shows no signs in CL of dissolution prior to
being overgrown by spilite-generation quartz. In contrast,
discordant embayments in the subtle growth zoning in the
spilite quartz shows that it was partly dissolved prior to
being overgrown by epidosite-generation quartz (as
observed in the lava-hosted epidosites, Section 5.1.2).
Epidosite alteration consisting of epidote + quartz +
titanite ± hematite ± rutile locally overprints spilite alter-
ation in the plagiogranites (Fig. S2). The epidosite minerals
are most obvious as fillings in the miarolitic cavities
(Fig. 8), overgrowing the spilite minerals where present.
Patchy epidosite alteration throughout the plagiogranite
matrix is common and is often accompanied by rectilinear
networks of cooling joints filled by epidote + quartz.
The paragenetic sequence of igneous and hydrother-
mal minerals in our samples of the axial plagiogranite
at Platanistasa, Troodos (Table 3), is essentially identical
to that of the Semail plagiogranites. Igneous quartz in
the rock matrix is graphically intergrown with plagioclase
(albitized by spilite alteration) and hosts isolated melt
inclusions. Precisely as observed in the Semail samples,
the quartz lining the miarolitic cavities exhibits four
stages of quartz overgrowths with characteristic textures
and CL intensities defining igneous, magmatic–hydrother-
mal, spilite and epidosite generations, with slight dissolu-
tion of quartz prior to precipitation of the epidosite
generation (Fig. S6).
Fig. 8. Field- and micro-photographs of epidotized post-axial plagiogranite at Bani Umar North, Semail ophiolite (Sample: LR17-BUN-Ta):
(A, B) Local patchy matrix epidotization and miarolitic cavities filled with Qz+Ep. (C) Euhedral Ep in miarolitic cavity with (D) primary LV
fluid inclusions along crystal growth zones. Abbreviations as in Figs. 5 and 6.
16 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
5.2. Characterization of fluid generations by fluid inclusion
petrography and microthermometry
A total of 77 thin- and thick-sections from 11 locations
in the Semail ophiolite and 10 sections from five locations
in the Troodos ophiolite were examined petrographically
for their fluid inclusions. Tables 2 and 3 show the types
of primary fluid inclusions found within their cogenetic host
minerals. The salinities and homogenization temperatures
are summarized in Tables 2 and 3, and detailed composi-
tional results are given in Table 4. The complete microther-
mometric data can be found in the Research Data.
In general, primary or pseudosecondary fluid inclusions
with tractable sizes are quite rare in the Semail spilites and
epidosites and only few of the many samples we collected
contain unequivocal petrographic relations that prove the
relative timing of inclusion trapping. Hydrothermal miner-
als that formed at higher temperatures, e.g., in cooling pla-
giogranites, tend to have the most abundant primary
inclusions. The fluid inclusion petrographic and microther-
mometric results pertaining to each generation are pre-
sented below in the order of their relative age with respect
to the local (outcrop-scale) timing of massive epidosites,
i.e., pre-, syn-, and post-epidosite.
5.2.1. Pre-epidosite melt inclusions in plagiogranites
In miarolitic cavities in the post-axial Semail plagiogran-
ites (e.g., Bani Umar North and Wadi Hawqayn), igneous
quartz hosts assemblages of primary inclusions consisting
of numerous birefringent crystallites with only minor aque-
ous vapor (u
vap
0.25) and aqueous liquid (u
liq
0.05)
(Fig. 9D1, D2). Based on Raman analysis, many of the
crystallites are silicates, including amphibole. The volumet-
ric proportions of the phases in the inclusions are uniform
within individual assemblages, indicating homogeneous
entrapment and implying that the solids are daughter min-
erals. The homogeneous entrapment and dominance of sil-
icate daughters indicate that the inclusions were initially
droplets of silicate melt. Following their entrapment in
quartz they exsolved a little aqueous liquid and vapor and
the remainder of the melt crystallized into the various
daughter minerals. These inclusions are labelled ‘‘melt
inclusionsin Table 2. The Platanistasa axial plagiogranite
in Troodos (e.g., sample LR18-Tr06) contains the same
kind of melt inclusions (Table 3).
5.2.2. Pre-epidosite, magmatic–hydrothermal fluid in
plagiogranites
In the Semail axial plagiogranite at Wadi Rajmi, locally
epidote-free miarolitic cavities contain magmatic–hy-
drothermal quartz that fills fractures in the earlier precipi-
tated igneous quartz. These healed fractures contain
pseudosecondary assemblages of very small (5mm diame-
ter) aqueous LVH inclusions, which represent the fluid that
precipitated the magmatic–hydrothermal quartz. Halite
dominates the inclusions (u
hal
0.4) with subordinate
vapor (u
vap
0.2), aqueous liquid (u
liq
0.3) and minor
amounts of hematite (presumably former magnetite trans-
formed by leakage of H
2
) and other unidentified solids.
The uniform phase proportions indicate homogeneous
entrapment of a dense hypersaline liquid. The inclusions
(Fig. S1F) tend to leak and decrepitate upon heating above
400 °C. No reliable data could be obtained by
microthermometry.
In the post-axial Semail plagiogranites (e.g., Lasail
South, Bani Umar North, Wadi Hawqayn), magmatic–
hydrothermal quartz in the miarolitic cavities contains
some primary but mostly pseudosecondary assemblages of
heterogeneously trapped inclusions, varying from a hyper-
saline LVH endmember to a vapor-rich LV endmember
(Fig. 9E1, E2). The hypersaline brines (u
vap
0.3; u
hal
-
0.3) show different sequences of phase transitions upon
heating. Some assemblages undergo halite melting via
LVH?LV at 210–365 °C prior to homogenization via
LV?L at 380–453 °C (e.g., in the Wadi Hawqayn and
Lasail South plagiogranites). Other assemblages show the
reverse order: LVH?LH at 231–270 °C and LH?Lat
355–420 °C. All these inclusions contain over 30 wt.%
NaCl
eq
, whereas the coexisting vapor-rich inclusions
(Fig. 9E2) contain 2.4–4.2 wt.% NaCl
eq
. Upon heating,
the vapor-rich inclusions show expansion of their bubbles
as they approach a dew-point transition (LV?V), but pre-
cise transition temperatures could not be measured owing
to migration of the L–V meniscus into the dark rims of
the inclusions.
In miarolitic cavities in the Platanistasa axial pla-
giogranite, Troodos (e.g., sample LR18-Tr06), magmatic–
hydrothermal quartz contains assemblages of primary,
LVH inclusions and coexisting LV inclusions with variable
vapor fractions distributed in 3D clusters (Fig. S6B–D).
Similar assemblages are present along healed fractures.
No clear endmember vapor-rich inclusions were found.
Microthermometry of LVH inclusions in two fluid inclu-
sion assemblages reveal T
m
(LVH?LV) of 290–330 °C (im-
plying 40 wt.% NaCl
eq
) and T
h
(LV?L) of 340–375 °C.
The Troodos inclusions are thus very similar to the Semail
inclusions.
5.2.3. Pre-epidosite spilite fluid in axial lavas and in
plagiogranites
Euhedral quartz in lava drainage cavities (e.g., in Geo-
times lavas at Rusays), which was locally overgrown by epi-
dote and quartz during later pervasive epidosite alteration,
hosts primary, homogeneously trapped assemblages of
aqueous fluid inclusions with u
vap
0.06. These inclusions
contain 2.4–2.6 wt.% NaCl
eq
and T
h
(LV?L) values are
128–140 °C. Presumably this pre-epidosite quartz belongs
to the spilite alteration assemblage.
In spilitized Geotimes lavas distant from any epidosite
or post-epidosite veins, the only amygdale quartz that we
found with tractable primary fluid inclusions occurs alone
without other coeval minerals, although it sits on earlier
chlorite. Whether the quartz belongs to the main albite +
chlorite + quartz spilite alteration or to the later
prehnite + quartz generation (see Supplementary Material
1) is therefore unclear solely from petrographic relations.
The euhedral quartz crystals in the amygdales host primary,
aqueous LV inclusions in homogeneously trapped assem-
blages (u
vap
is uniform within each assemblage, with values
for different samples and assemblages varying between 0.1
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 17
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Table 4
Summary of microthermometric analyses of fluid inclusions in this study.
Fluid generation Sample type Sample number Host
mineral
Fluid
inclusion
origin
Trapping
state
Phases
at T
lab
u
Vap
T
m
(Ice) or
T
m
(Halite)
1
[°C]
Salinity
[wt.% NaCl
eq
]
T
h
(LV?L) or
T
h
(LVH?LH)
2
[°C]
Pre-epidosite fluid generations
Magmatic–
Hydrothermal
Plagiogranite and
miarolitic cavities
LR17-BUN-Ta Qz P Hom LV+ solids 0.4 - - -
Qz PS Het LVH 0.3 355–373
1
48–50 263–273
2
LV 0.9 -1.6 to -1.4 2.6–2.9 -
Miarolitic cavities
in plagiogranite
LR18-WH-T02 Qz P Het LVH 0.3 210–223
1
32–33 380–410
LR18-Tr06* Qz PS Het LVH 0.3 290–330 39 340–375
LV 0.9 n.d. - n.d.
LR-TT06 Qz P Het LVH 0.3 340–365 43 422–453
LV 0.9 -2.5 to -2.2 4.0 n.d.
Spilite Miarolitic cavities
in plagiogranite
LR17-BUN-Ta Qz P Hom LV 0.6 -2.4 to -2.2 3.8–4.0 380–390
LR18-Tr06b* Qz P Hom LV 0.2 -3.5 to -2.9 4.8–5.7 318–323
Amygdales in lavas LR17-BU-Sp02 Qz P Hom LV 0.1 -1.9 to -1.8 3.1–3.2 160–180
AB16-4233 Qz P Hom LV 0.2 -2.0 to -1.8 3.2–3.4 265–275
Cavity in lavas LR17-F-Ep03b Qz P Hom LV 0.06 -1.5 to -1.4 2.4–2.6 130–140
Syn-epidosite fluid inclusion generation
Epidosite Massive epidosite in SD LR17-WH-SD-Ep03 Qz ? PEM/Het? LV 0.15 -3.1 to -2.8 4.7–5.1 200
V 0.8 -0.9 1.6 -
Qz S Hom LV 0.2 -1.8 3.1 250–260
Co-51* Qz P Hom LV 0.3 -1.9 3.2 277
Drainage cavities
in massive epidosite
in lavas
LR-A-Ep15-Qz2 Qz PS Hom LV 0.25 -1.7 to -1.5 2.6–2.9 280–295
LR-A-Ep15-Qz7 Qz PS Hom LV 0.25 -1.8 3.1 275–285
LR-A-Ep02a Qz PS Hom LV 0.25 -1.9 to -1.8 3.1–3.2 275–280
SG13-20-2-Qz2 Qz PS Hom LV 0.25 -1.5 2.6 280–290
SG13-20-2-Qz3 Qz PS Hom LV 0.3 -1.7 to -1.5 2.6–2.9 295–305
LR17-F-Ep03b Qz3 Qz PS Hom LV 0.25 -2.2 to -2.0 3.4–3.7 270–295
LR17-F-Ep03b Qz4 Qz P Hom LV 0.15 -2.3 to -2.0 3.4–3.9 220–240
LR18-Tr06* Qz P Hom LV 0.3 -3.1 to -3.0 5.0–5.1 380–390
Epidote–quartz veins LR17-WH-SD-Ep03 Ep P Hom LV 0.2 -1.9 to -1.7 2.9–3.2 250–260
SG13-41 Ep P Hom LV 0.3 -2.3 to -2.1 3.6–3.9 300–315
Epidotized miarolitic
cavities in plagiogranite
LR17-AB-T07b Ep P Hom LV 0.25 -2.4 to -2.3 3.9–4.0 365–370
LR18-Tr15* Ep P Hom LV 0.2 -2.3 to -2.2 3.6–3.9 358–369
LR18-Tr-T06* Ep P Hom LV 0.2 -3.1 to -3.0 5.0–5.1 275–280
LR17-BUN-Ta Qz P Hom LV 0.2 -1.9 to -1.8 3.1–3.2 252–256
Lava drainage cavity LR-AM-Ep06 Ep P Hom LV 0.25 -1.9 to -1.8 3.1–3.2 275–280
18 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
and 0.25; Fig. 6D). Salinities are 3.1–3.2 wt.% NaCl
eq
, with
T
h
(LV?L) of 160–180 °C in one sample and 265–270 °Cin
another.
In the axial plagiogranite at Wadi Rajmi, miarolitic cav-
ities that are locally epidote-free contain a spilite generation
of quartz (coating the early magmatic–hydrothermal
quartz) enclosing abundant submicroscopic needles of acti-
nolite and apatite (Fig. S1E). Intergrown with the needles
are primary assemblages of aqueous LV inclusions with
uniform u
vap
0.3. Salinities are 2.7–3.0 wt.% NaCl
eq
and T
h
(LV?L) values are 290–313 °C.
In the miarolitic cavities at the Bani Umar North post-
axial plagiogranite, spilite-generation quartz encloses crys-
tallites of actinolite and apatite (Fig. 9F1). Located between
the mineral inclusions are primary, euhedral, aqueous LV
inclusions with uniform u
vap
0.6 (Fig. 9F2), salinities of
3.7–4.0 wt.% NaCl
eq
and T
h
(LV?L) of 383–390 °C.
Similarly to the Semail plagiogranites, spilite-generation
quartz in the miarolitic cavities in the Platanistasa pla-
giogranite, Troodos, contains swarms of actinolite and
chlorite crystallites along growth zones (e.g., sample
LR18-Tr06; Fig. S6B). Between these mineral inclusions
are primary LV aqueous inclusions with uniform
u
vap
0.25, indicating homogeneous entrapment. The
salinity is 5.2 wt.% NaCl
eq
and T
h
(LV?L) varies over a
narrow range of 318–323 °C.
5.2.4. Epidosite fluid
In the Semail ophiolite, no primary fluid inclusions were
found within epidote, quartz or titanite in the rock matrix
of the massive, granoblastic epidosites that pervasively
replace former spilitized Geotimes basalts and dikes in the
SDC. However, tractable fluid inclusions were found in
three other settings: (1) in mm–cm long, freely grown epi-
dote and quartz crystals that line lava drainage cavities
within the massive epidosites (Fig. 7A), (2) epidosite-
generation quartz (very dark CL) in miarolitic cavities in
the plagiogranites (Fig. 9B, C), and (3) in mm–cm wide epi-
dote–quartz veins within the massive replacement epi-
dosites (Figs. 5B, 7B). Euhedral epidotes in these three
sample types contain unequivocal primary and pseudosec-
ondary assemblages of fluid inclusions, in which individual
inclusions are commonly elongated parallel to the long axis
of the host crystals (Fig. 7D). The assemblages consist of
aqueous LV inclusions (20 mm long) with uniform u
vap
of 0.25–0.3, indicating trapping of a homogeneous fluid.
Salinities are 3.1–4.0 wt.% NaCl
eq
and T
h
(LV?L) values
are 275–370 °C(Table 4), with individual assemblages dis-
playing narrow ranges of up to 10 degrees. In the coexisting
euhedral quartz, which shows clear petrographic evidence
of being coeval with epidote (e.g., Fig. 7E), fluid inclusion
assemblages are present on sets of healed fractures that ter-
minate inside the crystals at former growth zones. The
equant, anhedral LV inclusions (25 mm diameter) in these
pseudosecondary assemblages have constant u
vap
0.2–0.25 (Fig. 7F). Salinities are 2.6–3.9 wt.% NaCl
eq
and T
h
(LV?L) values are 252–305 °C(Table 4), with indi-
vidual assemblages exhibiting narrow ranges of less than 15
degrees.
Post-epidosite fluid generations
Quartz–hematite–
pyrite
Vein 260112-11A Qz P Hom LV 0.1 -1.6 to -1.5 2.4–2.7 144–145
LR-A-Ep17 Cal P Hom LV 0.05 -1.1 to -1 1.7–1.9 75–81
Prehnite–quartz Vein LR17-AH-03 Qz P?
(Isolated)
? LV ?? -2.2 to -2.0 3.4–3.7 220
Calcite Vein LR17-A-Cal02 Cal S Hom LV (or
mono-phase L)
0.05 -0.3
(mostly metastable)
0.21 62–78
LR17-WH-SD-Ep03 Qz P Hom LV 0.07 -0.4 to -0.2 0.4–0.7 56–72
Abbreviations: Cal – calcite; Ep – epidote; Qz – quartz; P – primary; PS – pseudosecondary, S – secondary; Hom – homogeneous; Het – heterogeneous; u
Vap
– vapor fraction; eq – equivalent; L –
liquid; V – vapor; H – halite. PEM – Post-entrapment modification.
For details see Supplementary Material 1; * – samples from the Troodo ophiolite.
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 19
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Outside the massive replacement epidosites in the Semail
lavas, vesicles and drainage cavities in the spilitized Geo-
times basalts occasionally contain fans of coarse-grained
epidote crystals without any accompanying quartz
(Fig. 10A, B). These epidotes may represent distal manifes-
tations of the epidotizing fluid that elsewhere formed
replacement epidosites. Individual epidote crystals contain
primary assemblages of aqueous LV inclusions arrayed
along growth zones (Fig. 10C, D). Uniform u
vap
0.25
indicates homogeneous entrapment. Salinities are 3.1–
3.2 wt.% NaCl
eq
and T
h
(LV?L) values are 275–280 °C.
In the Semail SDC, no primary fluid inclusions were
found in the granoblastic epidote within the massive
replacement epidosites (Fig. 5A). However, we found one
sample of these epidosites with anhedral quartz that con-
tains LV inclusions with variable u
vap
of 0.15–0.8. The
quartz displays no growth zones or healed fractures in the
CL image, and the inclusions appear to be distributed as
three-dimensional clusters within the host crystals. There-
fore, the inclusions are deduced to belong to assemblages
that are primary in origin. In the liquid-rich inclusions with
u
vap
0.15, salinity is 4.7–5.1 wt.% NaCl
eq
with
T
h
(LV?L) at 160–210 °C. The inclusions richest in vapor
(with u
vap
0.8) have a salinity of 1.6 wt.% NaCl
eq
. The
petrographic relationships are unclear, and in the available
assemblages it is not possible to determine whether the vari-
ation in u
vap
is due to post-entrapment modifications such
as necking-down, or whether it represents entrapment of a
heterogeneous fluid. The timing of entrapment with respect
to epidotization is also unclear: the anhedral quartz that
hosts the inclusions could have formed in either the spiliti-
zation or epidotization events. Owing to these uncertainties,
these inclusions are not considered further in this study.
We were unable to find unequivocal primary fluid inclu-
sions in epidote–quartz veins that cross-cut massive epi-
dosites in the Semail SDC (Fig. 5B). However, one vein
contains quartz crystals with secondary inclusions in healed
fractures (Fig. 5D), and the quartz is overgrown by much
Fig. 9. Microphotographs and relative timing relationships of different Qz generations and their inclusions in the post-axial Bani Umar North
plagiogranite, Semail ophiolite (Sample: LR17-BUN-Ta): (A) Transmitted light. (B) SEM-VPSE cathodoluminescence (CL): almost white –
igneous Qz with melt inclusions; medium-grey – magmatic–hydrothermal Qz in growth zones and along healed fractures in igneous Qz; dark
grey –hydrothermal Qz in latest growth zones and fracture fills. (C) Sketch of image B. (D1, D2) Primary melt inclusions in igneous Qz. (E)
Pseudosecondary LVH and coexisting V-dominated LV inclusions along growth zones in magmatic–hydrothermal Qz. (F) Primary LV
inclusions in growth zones of hydrothermal Qz coeval with actinolite (microphotograph from an equivalent thick section). Abbreviations: Act
– actinolite; H – halite; L – liquid; Qz – quartz; V – vapor.
20 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
younger calcite. The quartz-hosted inclusions have euhedral
to anhedral shapes and are 25 mm in diameter (Fig. 5D).
An assemblage of such inclusions (sample LR17-WH-SD-
Ep03) shows uniform u
vap
0.2, salinity of 2.9–3.2 wt.%
NaCl
eq
and T
h
(LV?L) at 250–260 °C. These values over-
lap with those found in primary inclusions in the lava drai-
nage cavities in the massive epidosites. In contrast, the late
calcite overgrowths in this sample contains primary inclu-
sions with notably lower salinity and T
h
than the quartz-
hosted inclusions (see Supplementary Material 1). It is
therefore probable that the secondary inclusions in quartz
trapped the fluid that formed the epidote–quartz veins.
In notable contrast to the Semail samples, our collection
of massive replacement epidosites from the Troodos SDC
contains quartz with abundant primary fluid inclusions,
distributed as 3D clusters throughout the crystals
(Fig. S5). These LV aqueous inclusions have elongated to
equant, subhedral to anhedral shapes and the assemblages
display uniform u
vap
0.3, indicating homogeneous
entrapment. Submicroscopic crystallites of epidote (con-
firmed by Raman analysis) are occasionally present within
the fluid inclusions. The high variation in u
epidote
(0–0.1)
rules out the origin of these crystallites as daughter crystals
and instead indicates that they were trapped simultaneously
with the homogeneous aqueous fluid. Salinities are 3.1–
3.2 wt.% NaCl
eq
and T
h
(LV?L) values cluster tightly at
272–283 °C.
In the Troodos plagiogranite at Platanistasa, epidote
that is intergrown with coeval quartz contains primary
assemblages of aqueous LV inclusions with uniform
Fig. 10. Field- and micro-photographs of a lava drainage cavity filled with euhedral epidote crystals: (A) Drainage cavity in the center of a
spilitized pillow lava filled with (B) euhedral epidote crystals up to 2 cm long. (C, D) Close-ups of euhedral epidote crystals hosting primary
LV inclusions. Abbreviations as in Figs. 5 and 6; Cal – calcite.
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 21
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
u
vap
0.2, indicating homogeneous entrapment. Values of
salinity are 3.7–3.9 wt.% NaCl
eq
, with T
h
(LV?L) at
358–369 °C.
5.2.5. Fluids in post-epidosite hydrothermal veins in the
Semail ophiolite
The post-epidosite prehnite–quartz veins, quartz–
hematite ± pyrite veins and calcite veins (Table 1) contain
primary or pseudosecondary fluid inclusions that yielded
salinities and homogenization temperatures as shown in
Fig. 3B and Tables 2 and 4 (details in Supplementary Mate-
rial 1). The calcite-generation fluids are easily distinguished
from the spilite and epidosite fluids, but the salinities and
T
h
values of the prehnite–quartz and quartz–hematite ± py
rite veins overlap with those of the spilite and epidosite flu-
ids (Fig. 3B). As we characterized the spilite and epidosite
fluids exclusively from their primary and pseudosecondary
inclusions, this overlap caused no confusion in identifica-
tion of the fluid generations.
5.3. Chemical compositions of the plagiogranite-, spilite- and
epidosite-generation fluids
One of the above-described quartz samples with
unequivocal petrographic evidence contains inclusions large
enough (20 mm diameter) for useful LA-ICP-MS analysis.
This sample from the Bani Umar North plagiogranite
(LR17-BUN-Ta; post-axial plagiogranite intruded the tran-
sition of the SDC to Geotimes lavas; Figs. 2 and 8) contains
the complete fluid sequence from igneous quartz with melt
inclusions, through magmatic–hydrothermal quartz with
hypersaline brine and vapor inclusions, spilite quartz host-
ing LV inclusions and finally epidosite quartz with LV
inclusions. Table 5 summarizes the analyses of latter three
fluid generations.
To characterize the magmatic–hydrothermal fluid, anal-
yses were made of 57 LVH inclusions and four V-rich inclu-
sions distributed among six fluid inclusion assemblages.
Raman analysis showed trace amounts of CO
2
vapor along
with H
2
O vapor in all the inclusions. For the LVH and
V-rich inclusions, LA-ICP-MS analysis yielded ratios of
the elements Li, B, Na, Mg, Cl, K, Ca, Fe, Br, Rb, Sr
and Ba. In the V-rich inclusions the elements Li, B, Ca,
Br and Ba were detected in only one inclusion. The domi-
nant cations in LVH and V-rich inclusions are Na, K and
Fe. Full analytical data including limits of detection
(LOD) can be found in the Research Data.
To characterize the spilite fluid, two assemblages con-
taining a total of 28 inclusions were analyzed by Raman
spectroscopy, again showing traces of CO
2
, and by LA-
ICP-MS, yielding ratios of all 12 elements. Calculations
according to Section 5.3 (Fig. S12) show that the observed
range of T
m
(Ice) from -2.4 to -1.4 °C corresponds to 2.7–
4.4 wt.% TDS.
In the vapor bubbles of the fluid inclusions in the
epidosite-generation quartz, H
2
O was the only component
detected by Raman spectroscopy. Analysis by LA-ICP-
MS of 27 LV inclusions in two assemblages yielded ratios
of all elements except for Fe (Research Data). Applying
the same calculations to the epidosite fluids as for the spilite
fluids, the range of salinity is 2.5–4.3 wt.% TDS, essentially
identical to that of the spilite fluid.
In the spilite and epidosite fluids, Na and Cl are the most
abundant elements followed by Ca, whereas in the hyper-
saline brines and in the coexisting magmatic–hydrothermal
vapors K and Fe exceed Ca. For the epidosite fluid, only a
maximum possible concentration of Fe is available, set by
the lowest LOD in the analyzed assemblage
(7.0 10
-4
mol/kg
H2O
). This LOD concentration is far
too high for equilibrium with the hematite-bearing epi-
dosite mineral assemblage at the estimated fluid trapping
temperature (275–295 °C; Section 5.4). We therefore set
the Fe concentration to the fully speciated, thermodynam-
ically calculated equilibrium value at 300 °C(Weber
et al., 2021), which is about 4.0 10
-6
mol/kg
H2O
.
5.4. P–T conditions of fluid inclusion trapping
Since the spilitizing and epidotizing fluids were trapped
as single-phase fluids in the Semail inclusions, their homog-
enization temperatures represent minimum limits on fluid
trapping temperatures (T
trap
; see principles in Diamond,
2003). The true values of T
trap
and of trapping pressure
(P
trap
) lie somewhere along the P–T trajectories of the fluid
inclusion isochores (lines of constant density). Fig. 11
shows the range of isochores at the investigated localities
calculated from salinity and T
h
(LV?L) values (e.g., point
1inFig. 11A) using the AqSo_NaCl P–T–V
m
–x correlation
for the binary H
2
O–NaCl system in Bakker (2018; based on
Driesner and Heinrich, 2007; Driesner, 2007).
The isochores in Fig. 11 can be used to estimate the pres-
sure of fluid trapping (P
trap
) by first estimating the depth of
trapping below sealevel (i.e., the sum of seawater depth plus
paleostratigraphic depth below seafloor) and then adopting
a relationship between depth and fluid pressure. We have
used the synvolcanic ocean depth of 3.1–3.5 km (corre-
sponding to pressures of 31–35 MPa) defined by the overlap
between sedimentary CCD evidence and volatile contents in
glass in the paleobathymetric study of the Semail ophiolite
by Belgrano et al. (2021). The paleostratigraphic depth of
each sample was converted to hydrostatic pressure taking
into account the density of the fluid at trapping, as defined
by the relevant isochores. Details are provided in Supple-
mentary Material 2.
The maximum and minimum P
trap
values can be used
to determine the temperature of trapping (T
trap
) and
hence the temperature of rock alteration, by intersection
with the isochores in Fig. 11. For example, the range
in P
trap
for the spilite fluid at the Rusays locality is
31–54 MPa. In Fig. 11A these bounding pressures (blue
P
trap
lines) intersect isochore 1 at 146 °C (point 2) and
at 159 °C (point 3). The small spread in isochores con-
strains the range of T
trap
to 25 degrees, i.e., from 146–
171 °C, thereby defining a field (green) of possible P–T
conditions of fluid trapping. In summary, the Semail
localities record spilite alteration over a broad, 290
degree range of 146–438 °C at fluid pressures of >31–
54 (Fig. 11A), whereas epidosite alteration is recorded
at a narrower, 180 degree range of 255–436 °C at fluid
pressures of 35–68 MPa (Fig. 11B).
22 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Table 5
Average composition of spilite fluid and epidosite fluid compared to magmatic–hydrothermal hypersaline brine and coexisting vapor from a plagiogranite (this study, sample LR17-BUN-Ta, Bani
Umar North) and compared to modern seafloor hydrothermal vent fluid and unmodified seawater.
Epidosite fluid
(27 inclusions,
3.2 wt.% TDS
1
)
1 SD Spilite fluid
(28 inclusions,
4.2 wt.% TDS
1
)
1 SD Plagiogranite
hypersaline brine
(57 inclusions
2
,
40.7 wt.% TDS
1
)
1 SD Plagiogranite vapor
(4 inclusions
2
,
2.9 wt.% TDS
1
)
1 SD Seawater
3
(3.5 wt.% TDS
1
)
T°C 274–294
401–438
430–300 430–300 0
Li mol/kg
H2O
0.005 0.004 3.1E-4 4.7E-5 0.003 0.003 2.3E-4 2.8E-5
B mol/kg
H2O
6.6E-4 2.5E-4 9.3E-4 3.3E-4 0.002 0.001 8.9E-5 4.2E-4
Na mol/kg
H2O
0.439 – 0.396 – 4.195 – 0.231 0.486
Mg mol/kg
H2O
9.8eE-4 0.001 0.001 0.003 0.002 0.008 3.9E6 0.055
Cl mol/kg
H2O
0.599 * – 0.727 * – 8.20 0.525 * 0.566
K mol/kg
H2O
0.012 0.004 0.056 0.035 1.68 0.305 0.075 0.014 0.011
Ca mol/kg
H2O
0.069 0.021 0.094 0.072 0.245 0.079 0.009 0.011
Fe mol/kg
H2O
4.0E-6
(Calc.)
0.029 0.023 0.989 0.376 0.082 0.022 3.4E-9
Rb mol/kg
H2O
1.3E-5 7.6E-6 7.6E-5 3.9E-5 0.002 0.001 2.0E-4 1.8E-4 1.3E-6
Sr mol/kg
H2O
1.3E-4 6.0E-5 9.0E-5 4.1E-5 0.001 0.001 4.2E-5 3.6E-5 9.4E-5
Br mol/kg
H2O
0.002 0.001 0.002 0.002 0.032 0.004 8.7E-4
Ba mol/kg
H2O
4.9E-5 2.1E-5 5.8E-5 3.2E-5 0.034 0.025 4.7E-5 1.4E-7
CO
2
mol/kg
H2O
n.d. – trace – trace – trace
1
Total dissolved solids (not NaCl equivalent);
2
hypersaline liquid in phase-separated LVH inclusions;
3
Na, Mg, Cl, Ca, Sr from Millero et al., (2008); Li, B, Fe and Br from Butterfield et al.
(1990); Rb in vent fluid from 13°N East Pacific Rise and in seawater from Palmer and Edmond (1989); Ba from Charlou et al. (1996);
Temperature of fluid trapping estimated from reconstructed
hydrostatic pressure (see text for explanation); *calculated from charge balance assuming Cl is the only significant anion; (Calc.) – from Weber et al., 2021; Mg in the vent fluid corresponds to the
limit of effective detection owing to contamination by seawater; n.d. – Not detected.
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 23
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
6. DISCUSSION
6.1. Reliability and representativeness of results
In all our analyzed assemblages of homogeneously
trapped fluid inclusions with low salinities (<5.3 wt.%
NaCl), the ranges of T
m
(Ice) are less than 0.3 degrees and
the ranges of T
h
are typically less than 20 degrees. This tight
grouping of values within individual assemblages supports
the conclusion that none of the inclusions selected for anal-
ysis has undergone post-entrapment modification. Further,
all the fluid inclusions that we have observed, regardless of
whether they have been modified after entrapment and
hence ignored during analysis, fall into one of the six aque-
ous fluid generations defined by our petrography (Table 2).
However, only a small number of tractable assemblages
of unequivocal primary or pseudosecondary fluid inclusions
could be used to characterize the fluid generations. Indeed,
none of the Semail massive epidosite samples yielded pri-
mary fluid inclusion assemblages in the rock matrix, and
so our deductions rely on assemblages in the cogenetic
and spatially associated cavity fillings and epidote–quartz
veins. Despite having only a limited number of unequivocal
fluid inclusion assemblages for analysis in the Semail sam-
ples, the fact that we have examined more than 75 petro-
graphic sections lends confidence that our results are
representative, as well as pertinent for the alteration pro-
cesses of interest.
6.2. Origin of the plagiogranite fluids
The miarolitic cavities in the axial and post-axial pla-
giogranites appear texturally identical, although they are
somewhat larger and more abundant in the post-axial
stocks. Two field features have important genetic implica-
tions: (1) the amphibole-free pegmatitic intergrowths of
coarse igneous quartz and plagioclase around the cavities
are indicative of enhanced water contents in the most
evolved portions of the magma; (2) the myriads of mostly
unconnected cavities are regularly distributed throughout
large volumes of the plagiogranites. These features attest
to the origin of the cavities as pockets of magmatic–hy-
drothermal fluids that exsolved from the melt during the
final stage of plagiogranite crystallization.
6.3. Constraints on the origin of the spilite fluid
6.3.1. Endmember vs. mixed ‘‘spilitefluid in Bani Umar
North plagiogranite
The spilite-generation fluid inclusions in the Bani Umar
North plagiogranite are clearly later than the hypersaline
liquid + low-salinity vapor inclusions in the magmatic–hy-
drothermal quartz (Section 5.2). Mixing calculations based
on the multi-element compositions of the two fluid genera-
tions (Supplementary Material 3) rule out the possibility
that the analyzed spilite fluid is contaminated by mag-
matic–hydrothermal hypersaline liquid or low-salinity
vapor. This spilite fluid is therefore taken as an example
of the fluid that induced widespread spilite alteration in
the sheeted dikes and lavas (with allowance for differences
in local water–rock ratios and P–T conditions of spilitiza-
tion subtly modifying its element ratios).
6.3.2. Phase state and origin of salinity in the spilite fluid
Our samples from the Semail ophiolite show that the
axial Geotimes lavas and the axial and post-axial pla-
giogranites were spilitized by circulation of single-phase
(liquid) aqueous solutions with salinities of 2.4–5.7 wt.%
NaCl
eq
(Fig. 12), straddling the salinity range of modern
seawater (2.8–3.6 wt.% NaCl
eq
). Moreover, the Bani Umar
North spilite fluid inclusions have a molar Br/Cl ratio
Fig. 11. P–T conditions of spilite and epidosite alteration at the
studied locations in the Semail ophiolite, illustrated in the P–T
diagram of seawater modelled with 3.2 wt.% NaCl
eq
. Thin lines
represent isochores through the liquid field calculated using the
correlation of Bakker (2018) for the ranges of T
h
and salinities
determined by microthermometry at each locality. Fluid P
min
and
P
max
are calculated as explained in Supplementary Material 2.
Intersection of the isochores with P
min
and P
max
defines the
possible P–T conditions of rock alteration (colored bands) at each
site. (A) Spilite alteration. (B) Epidosite alteration.
24 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
(0.002 ± 0.002 SD) that is compatible within the large ana-
lytical uncertainty with the molar Br/Cl ratio of modern
seawater (0.0015; Millero et al., 2008; Butterfield et al.,
1990). Based on the preservation of this ratio of conserva-
tive elements, an origin of the spilite fluid as modified sea-
water appears certain. In a third of the samples, the salinity
is between 3.1 and 3.2 wt.% NaCl
eq
, which we interpret to
represent the chlorinity of Cenomanian seawater in the
Semail realm. The broader salinity range of the remaining
samples (well outside the error of our T
m
(Ice) measure-
ments; Fig. 12) is similar to but smaller than that found
in spilite fluids elsewhere (Fig. 3A, field 1). Possible causes
for this spread have been discussed in the literature and
their applicability to our results is examined in the
following.
(a) Some researchers have suggested that the salinity of
Cretaceous seawater was more variable than that of
modern seawater (e.g., Brass et al., 1982; Hay,
2008; Friedrich et al., 2008) but others argue for con-
stant salinities during the Phanerozoic (e.g., Johnson
and Goldstein, 1993), hence the issue is unclear. On
the other hand, it is implausible that seawater above
the Semail crust could have evolved appreciably in
salinity during the relatively short, 1 million year
duration of volcanic activity and subsequent cooling.
(b) Rock hydration during the hydrothermal alteration
has often been suggested as a way that salinity might
increase (e.g., Delaney et al., 1987), but in view of the
elevated water–rock ratios implied by the spilite alter-
ation, this process is unlikely to have increased the
salinity of the fluid measurably.
(c) As far as decreasing the chlorinity is concerned,
amphibole with 30–16000 mg/g Cl (Kendrick,
2019), chlorite with 20–700 mg/g Cl (Kendrick,
2019) and clay minerals, e.g., celadonite with
140–4000 mg/g Cl (Kendrick, 2019) are the only pro-
ven sinks for chloride among the alteration minerals
in the crust at depths up to amphibolite-facies condi-
tions. However, owing to the elevated water–rock
ratios, their cumulative effect on the fluid salinity is
also likely to have been minimal.
(d) Studies of active seafloor vent fluids have revealed a
wider range in salinities than we have found in the
Semail spilite fluids. In fact, vent fluids almost always
have chlorinities higher or lower than seawater
(German and Von Damm, 2006), varying from 0.2
to 5.2 wt.% NaCl
eq
(Von Damm, 2000; Gallant and
Von Damm, 2006). This range, and the even wider
spread found in spilite fluid inclusions in the in-situ
crust (<0.1–10 wt.% NaCl
eq
;Fig. 1A), is widely
attributed to sub-seafloor boiling near the venting
Fig. 12. Salinities of spilite- and epidosite-fluids from the Troodos (labelled Tr) and Semail ophiolites measured in this study. Range bars
show the variation of salinity in individual fluid inclusion assemblages. Instrumental uncertainty (±0.18 wt.% NaCl
eq
= ±0.1 °C) is not
shown. Modern seawater after Levitus et al. (1994). Range of modern seafloor vent fluids from Von Damm (2000) and Gallant and Von
Damm (2006).
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 25
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
sites or to liquid–vapor exsolution (i.e., ‘‘phase sepa-
rationvia brine condensation) and segregation with
or without remixing deep in the crust (e.g., Delaney
et al., 1987; German and Von Damm, 2006).
The similarity between some of our salinity data and
those from vent fluids (Fig. 12) calls for a closer examina-
tion of our evidence for fluid inclusion trapping in the
single-phase state. Following the standard interpretation
of the trapping mechanism of fluid inclusion assemblages
(e.g., Diamond, 2003), the absence of vapor-rich inclusions
in the spilite-generation assemblages implies merely that
liquid was locally present alone during fluid inclusion
entrapment. In principle, it could be argued that a vapor
phase did in fact coexist with the liquid in the spilite rock
pores, but that it was not trapped in the hydrothermal
quartz owing to its different surface wetting and transport
properties. However, quartz is known to readily trap vapor
as well as liquid in a wide range of hydrothermal environ-
ments (the hypersaline liquid and vapor inclusions in the
Semail plagiogranites being a case in point). To our knowl-
edge, the literature on experimental and observable modern
environments (e.g., geothermal wells) contains no cases of
demonstrated liquid–vapor coexistence in which quartz
has systematically trapped liquid but excluded vapor. Since
we have examined numerous samples throughout the
Semail ophiolite and failed to observe vapor, we conclude
that the spilitizing fluid was not in contact with vapor at
the time of fluid inclusion entrapment. The liquid–vapor
exsolution that occasionally shifted its salinity must have
occurred previously along the upflow path at greater depth
in the crust.
The changes in salinity of our analyzed spilite fluids can
be tracked by reconstructing their possible deep flow paths,
based on the known P
trap
T
trap
conditions, the P–T loci of
phase boundaries of seawater and the P–T dependency of
quartz solubility. Fig. 13 schematically illustrates deep
sub-horizontal flow along the base of the Semail SDC near
an active magma chamber (not shown). Since entrapment
of primary fluid inclusions requires precipitation of new
quartz, the disposition of quartz solubility isopleths
(Akinfiev and Diamond, 2009) constrains the flow path to
intersect the inclusion trapping conditions (green fields)
during upflow and cooling. Fig. 13A shows the example
of a spilite fluid with seawater chlorinity (3.2 wt.% NaCl
eq
)
trapped at the Highway 7 locality. The fluid remains in the
liquid state with fixed chlorinity throughout its flow path,
becoming trapped in inclusions at point 1.
Drawing on Delaney et al. (1987), a mechanism of gen-
erating spilite fluids with the observed range of salinities
between 2.5 and 3.9 wt.% NaCl
eq
via liquid–vapor separa-
tion is illustrated in Fig. 13B. Along the indicated horizon-
tal and initial upflow segments, the fluid expands
continuously, changing from a dense liquid to a vapor-
like state at temperatures above the critical isochore (dotted
line). At point 1 (arbitrarily chosen at 53 MPa, 495 °C) the
fluid intersects the dew curve (i.e., the segment of the LV
curve at T>T
critical
), inducing condensation of small
amounts of dense, high-salinity liquid. Decompression
within the LV field by just 2 MPa and cooling by 5 °C
causes exsolution of 6 vol.% hypersaline liquid (28 wt.%
NaCl
eq
), reducing the salinity of the dominant vapor to
2.5 wt.% NaCl (point 2 at 51 MPa, 490 °C). Since the phase
boundaries for the coexisting fluids lie very close to the dew
curve of seawater, further upflow brings both fluids back
into the one-phase field, although presumably the dense liq-
uid lags behind owing to the contrast in transport proper-
ties. At point 3 the dominant vapor may mix with
residual hypersaline liquid from previous events of liquid–
vapor separation, and so attain a higher salinity, e.g.
3.9 wt.% NaCl
eq
, as in the quartz-hosted spilite fluid at Bani
Umar North (green field). Continued upflow, constrained
in Fig. 13B to follow a path of decreasing quartz solubility,
leads to precipitation of quartz and trapping of primary
fluid inclusions at point 4. Thereafter the fluid is shown
ascending to the seafloor and venting at 31 MPa and
380 °C. A case where the low-salinity ‘‘vaporgenerated
at point 2 in Fig. 13A has not mixed with hypersaline liq-
uid, but simply contracted into the liquid state upon
upflow, is given by the spilite-type fluid inclusions with
salinity of 2.5 wt.% NaCl
eq
at Rusays (Fig. 11A).
Although constrained by our observations, much of the
flow path in Fig. 13B is arbitrary, notably the cooling
between points 2 to 3. Nevertheless, the path appears feasi-
ble because it is similar to the dew-curve intersection
deduced for fluids in the 4.5 km IDDP-2 borehole into
the sheeted dike complex below Iceland (Bali et al., 2020).
We see no obstacle to the fluid at point 1 in Fig. 13B having
penetrated the LV field under open-system conditions, as
segregation of the dense liquid lowers the initial fluid vol-
ume by 2 vol.%. Irrespective of the liquid condensation,
however, the illustrated upflow path of the ‘‘vaporis char-
acterized by significant volume expansion.
6.3.3. Validity of the spilite fluid composition
As far as we are aware, the multi-element analyses of the
spilite and epidosite fluids presented in Section 5.3 are the
first obtained on fluid inclusions of near-seawater salinity
from oceanic crust. Although measured by state-of-the-art
LA-ICPMS, they are relatively imprecise; at least an order
or magnitude less precise than typical analyses of seafloor
vents. This is due to the technical challenges of analyzing
picogram quantities of solutes in such small and weakly sal-
ine fluid inclusions. On the other hand, the analyses do not
suffer from contamination by cold seawater, as is the case
for most sampled vent fluids.
To test the plausibility and broad accuracy of the spilite
fluid analyses, two tests can be conducted. The first is a qual-
itative check that elements depleted in the spilite rocks by
seawater–basalt interaction are enriched in the spilite fluid
and visa versa, thereby respecting crustal-scale mass-
balance constraints (Fig. 14). Relative to the mean composi-
tion of four fresh Geotimes–Lasail basalt glasses, the spili-
tized Geotimes lavas in the Semail ophiolite are depleted in
K, Ca, Fe, Rb, Sr and Ba. As expected, the spilite fluid is
enriched in these elements relative to seawater. Most notable
is the huge enrichment of Fe in the spilite fluid. Sodium and
Mg, in contrast, are enriched in the spilite rocks relative to
fresh basalt and are accordingly depleted in the spilite fluid
relative to seawater. Lithium and B do not appear to follow
26 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
Fig. 13. Schematic P–T paths illustrating scenarios of trapping of spilite fluids in the analyzed fluid inclusions. Similar scenarios apply to the
epidosite fluids. Stratigraphic columns pertain to the specific localities at the times of trapping. Isopleths of quartz solubility (mmol/kg
H2O
)
calculated from Akinfiev and Diamond (2009) using molar volumes from AqSo_NaCl (Bakker, 2018). (A) Highway 7 locality, sample AB16-
4233: fluid with seawater salinity (3.2 wt.% NaCl
eq
) remains in the liquid field and is trapped as primary fluid inclusions in quartz at point 1.
(B) Bani Umar North locality, sample LR17-BUN-Ta: liquid initially with seawater salinity expands into a vapor-like fluid along its upflow
path and intersects the dew curve at point 1. At point 2 it exsolves minor (6 vol.%) hypersaline liquid with salinity of 28 wt.% NaCl
eq
, shifting
the salinity of the remaining vapor to 2.5 wt.% NaCl. Cooling of the two fluids into the one-phase field allows partial mixing at point 3,
generating the 3.9 wt.% NaCl fluid trapped as primary fluid inclusions in quartz at point 4.
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 27
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
this simple scheme. Apart from the latter exceptions, the
composition of the spilite fluid is consistent with the mass-
balance constraints, supporting the validity of the analyses.
A second test is a qualitative check on whether the dif-
ferences in element concentrations between the spilite fluid
and seawater match those expected for seawater–basalt
interaction up to greenschist-facies conditions (Fig. 15).
The extreme depletion of the fluid in Mg conforms to the
experimentally verified metasomatic trends of heated sea-
water reacting with basalt (e.g., Mottl, 1983). In the absence
of major mineral sinks for B and Br in the altered basalts,
the concentrations of these elements lie close to those of
seawater, as expected. Owing to the stabilization of albite
in the spilite mineral assemblage in exchange for anorthite
component in plagioclase, Na is present at only about
80% of its average molar concentration in seawater. Cal-
cium is enriched relative to Cenomanian seawater, but the
average Sr concentration is indistinguishable from that in
modern seawater, suggesting that their ratio has been mod-
ified by water–rock interaction. Along with Ca, the concen-
trations of fluid-mobile elements Li, K, Rb and Ba are all
highly enriched in the spilite fluid. Iron, the only redox ele-
ment plotted, is significantly enriched owing to the reduced
redox state of the spilite fluid compared to that of seawater.
Overall, the fit of the element concentrations with the
expectations underscores the validity of the analyses.
6.4. Constraints on the origin of the epidosite fluid
6.4.1. Salinity, phase state and relationship of the epidosite
fluid to seawater
All our analyses of epidosite-generation fluids from the
Semail and Troodos ophiolites indicate that they were
single-phase aqueous liquids at the moment of trapping in
fluid inclusions. Their salinities vary from 2.6 to 5.1 wt.%
NaCl
eq
, i.e., from values below to above the range of Ceno-
manian seawater (Fig. 12). Based on the same reasoning as
in Section 6.3.2, high-temperature liquid–vapor exsolution
at depth below the sampling sites (e.g. Fig. 13B) remains
the most plausible explanation for the range in salinities.
Again, the implication is that the epidosite fluids were
trapped along upflow paths. Nevertheless, numerous fluid
inclusion samples in Fig. 12 cluster at 3.1–3.2 wt.% NaCl
eq
.
This value could well represent the chlorinity (560 mmol/
kg Cl) of Cenomanian seawater that had not undergone
any liquid–vapor exsolution during the hydrothermal alter-
ation in the Semail and Troodos ophiolites (e.g. Fig. 13A).
Our salinity and T
h
measurements agree with the subset
of previous studies that attribute epidosite formation to
seawater-like fluids, as summarized in Fig. 3B and
Section 3.2 (Richardson et al., 1987; Nehlig and Juteau,
1988; Schiffman and Smith, 1988) but they contradict the
second subset that attribute formation of massive replace-
Fig. 14. Enrichment and depletion of elements in spilite fluid (this study) relative to modern seawater
1
(Palmer and Edmond, 1989; Butterfield
et al., 1990; Charlou et al., 1996; Millero et al., 2008), and in bulk spilitized Geotimes and Lasail lavas
2
(Belgrano and Diamond, 2019, with Ca
corrected for contamination by post-calcite veinlets using 61 own CO
2
analyses) relative to fresh basalt glass
3
(V1 = Geotimes/Lasail basalt,
Kusano et al., 2017).
28 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
ment epidosites to hypersaline brines (Cowan and Cann,
1988; Juteau et al., 2000). In agreement with the observa-
tions of Kelley et al. (1992) in Troodos, we have found that
hypersaline brines are spatially and temporally restricted to
plagiogranites in both the Semail and Troodos ophiolites
and that the brines in fact exsolved from the plagiogranite
melts. The contradicted studies of Cowan and Cann
(1988) and Juteau et al. (2000) drew their evidence for
hypersaline brines from epidotized plagiogranites in Troo-
dos and Semail, respectively. In contrast to the findings in
the massive epidosites, Kelley et al. (1992) found hyper-
saline and coexisting vapor inclusions in cm-size epidote
pods in the Troodos plagiogranites and suggested that the
epidote formed by deuteric alteration. We do not rule out
this suggestion, although we have not found such inclusions
in the epidote + quartz assemblages in our Troodos sam-
ples. However, our CL imaging has shown that the hyper-
saline brines are older than and genetically unrelated to
the overprinting massive epidosites. Furthermore, spilite
alteration systematically occurs in the time interval between
magmatic fluid exsolution and late epidosite alteration in
the plagiogranites of both ophiolites. Since the zonation
of the quartz in the plagiogranites is complex and extremely
difficult to discern under the normal petrographic micro-
scope, we suspect that the lack of CL imaging in the earlier,
contradicted studies led to misleading petrographic obser-
vations. Thus, all our fluid inclusion evidence speaks
against the formation of large massive replacement epi-
dosites by magmatic–hydrothermal hypersaline fluids.
However, recent REE and Sr-isotope studies (Anenburg
et al., 2015; Fox et al., 2020) argue that plagiogranite fluids
play the prime role in forming massive epidosites in both
the plagiogranites and the surrounding SDC. Further work
is required to reconcile the two types of evidence.
6.4.2. Chemical composition of the epidosite fluid
Fig. 15 and Table 5 show the elemental composition of
the reconstructed epidosite fluid trapped at 275–295 °Cat
Bani Umar North, which has the unmodified salinity of
Cenomanian seawater. The pattern of element concentra-
tions in Fig. 15 is in many respects similar to that of the spi-
lite fluid, hence we consider the LA-ICP-MS analyses to be
just as valid as those of the spilite fluid (Section 6.3.3). As
epidosites in Oman are known to replace precursor spilites
(Section 2.3), one way to view the epidosite fluid is as an
evolved form of an earlier spilite fluid that was chemically
modified by fluid–rock interaction (but not necessarily by
liquid–vapor exsolution), such that Na and Mg can be
Fig. 15. Element concentrations (molality) in spilite fluid at 425 °C (this study) and epidosite fluid at 285 °C (this study) compared to modern
seawater (Palmer and Edmond, 1989; Butterfield et al., 1990; Charlou et al., 1996; Millero et al., 2008), Cenomanian seawater (Timofeeff et al.,
2006) and modern seafloor vent fluids (range of vent fluids: Juan de Fuca, Butterfield et al., 1990; Scotia Ridge vents, James et al., 2014; TAG,
Schmidt et al., 2017 and references therein). LoD: Limit of detection. Calc: Thermodynamically calculated concentration of Fe from Weber
et al. (2021), see text.
L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx 29
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
quantitatively leached and replaced by Ca (e.g., Weber
et al., 2021). However, the Ca/Mg and Ca/Na ratios in
the analyzed epidosite and spilite fluids are not significantly
different given the available precision. Potassium and Rb
appear to be lower in the epidosite fluid, perhaps because
these elements had already been depleted in the spilite pre-
cursors. However, the most obvious difference between the
two fluids is the concentration of Fe. The very low calcu-
lated content in the epidosite fluid (4.0 10
-6
mol/kg
H2O
)
is consistent with all Fe in the epidosite alteration being oxi-
dized and therefore insoluble. Finally, the conservative ele-
ments B and Br (and Cl) appear similar, as expected from
the lack of mineral sinks for these elements.
6.4.3. Relationship of spilite and epidosite fluids to vent fluids
on the modern basaltic seafloor
The spilite and epidosite fluids are compared to modern
basalt-hosted, black-smoker seafloor vent fluids in Fig. 15.
Mean values of the East Scotia Ridge vent (James et al.,
2014), Juan de Fuca (Butterfield et al., 1990) and TAG
(Schmidt et al., 2017 and references therein) have been cho-
sen for this comparison because they include a large num-
ber of analyzed elements. The spilite fluid matches several
features of the vent fluids (contents of Li, B, Na, Cl, K,
Br and Sr), but Ca, Fe (and possibly Rb and Ba) are higher
in the spilite fluid. The elevated Ca and Fe concentrations
may reflect the fact that these elements (and Ba) tend to
precipitate as sulfates and sulfides just below or at the sea-
floor, such that typical vent-sampling methods do not cap-
ture the higher concentrations present in the deeper
upflowing fluid. Magnesium in the spilite fluid is notably
lower than in the plotted mean vent fluid (Mg value from
East Scotia Ridge vents), but this may be due to the difficul-
ties of determining Mg concentration in the pristine venting
fluids, owing to contamination by cold seawater (German
and Von Damm, 2006). Based on these comparisons, it
seems plausible that the analyzed spilite fluid at 425 °C
could have acted as a feeder fluid to a typical basalt-
hosted, black-smoker vent.
The comments made above regarding B, Mg, Br, Rb and
Sr in the spilite fluid compared to the vent fluids also apply
to the epidosite fluid. However, as emphasized by Weber
et al. (2021), the estimated concentration of Fe in the epi-
dosite fluid is incompatible with black-smoker type fluids.
Although the epidosite fluids represent upflowing
hydrothermal solutions, they appear to be too poor in Fe
to form pyrite-rich VMS deposits at or just below the sea-
floor. We are unaware of an other type of seafloor vent fluid
that could match our analyzed epidosite fluid. While their
elevated Ca content suggests that the epidosite fluids would
generate sulphate chimneys upon focused venting into nor-
mal sulphate-rich seawater, it is also possible that they vent
diffusely without noticeable seafloor precipitates and have
therefore not been recognized so far.
7. SUMMARY AND CONCLUSIONS
We have conducted a field, petrographic and analytical
study on fluid inclusions petrographically attributable to
spilite-type (chlorite + albite + quartz ± actinolite ± epidote)
and epidosite-type (epidote + quartz + titanite + hematite or
magnetite) hydrothermal alteration in the basaltic upper
crust of the Semail ophiolite, with supporting samples from
the Troodos ophiolite. The sampled geological settings in
the Semail ophiolite include the sheeted dike complex, the
comagmatic (axial) Geotimes lavas, and axial and post-
axial plagiogranite stocks in the sheeted dikes and Geotimes
lavas.
All our samples consistently show that the spilite and
epidosite alteration was caused by single-phase (homoge-
neous) aqueous liquids. However, their salinities vary
between sampling sites. In a third of the samples, the salin-
ity is between 3.1 and 3.2 wt.% NaCl
eq
, which we interpret
to represent the chlorinity of Cenomanian seawater in the
Semail realm. The remaining samples have salinities as
low as 2.4 wt.% NaCl
eq
and as high as 5.7 wt.% NaCl
eq
.
This spread can be quantitatively explained by modest
degrees of liquid–vapor separation and remixing that
occurred deeper in the crust prior to the fluids being
trapped as single-phase liquids at the sampling sites. The
implication is that samples with salinities lower or higher
than Cenomanian seawater were formed along upflow seg-
ments of the fluid paths, consistent with previous studies of
spilite-related fluid inclusions in in-situ crust and in the
Semail and Troodos ophiolites.
Similarly, our results regarding the phase state and salin-
ity of the epidosite fluid agree with several previous studies
on the Semail and Troodos ophiolites, whereas other stud-
ies have argued that hypersaline brines formed the massive
replacement epidosite bodies. According to our sampling in
the two ophiolites, hypersaline brines, often accompanied
by vapor, occur only in plagiogranites (both axial and
post-axial). These magmatic–hydrothermal fluids exsolved
from the plagiogranite melts during final crystallization.
Our CL imaging of plagiogranite samples from both ophi-
olites revealed that the magmatic–hydrothermal fluids pre-
date and are genetically unrelated to the younger spilite
fluid and even younger epidosite fluid.
Consideration of the Semail volcanostratigraphy con-
strains the examined sites to maximum depths between
1470 m and 3600 m below seafloor at the time of fluid inclu-
sion trapping. Taking account of paleo-ocean depths
between 3100 and 3500 m, the spilite fluids were trapped
at 145–440 °C and >31–54 MPa and the epidosite fluids
at 255–435 °C and 35–68 MPa.
The precision of our elemental analyses of the spilite
and epidosite fluids is well below that of typical vent fluid
analyses, but they nevertheless yield valid and useful
results. The Br/Cl ratio of the spilite fluid matches that
of modern seawater, and the concentrations of most of
the other analyzed elements in the spilite fluid match those
of modern basalt-hosted black-smoker vents, including
very low Mg. The exceptions are elements that typically
precipitate just below or at the seafloor, prior to sampling
of vent fluids, namely Ca and Fe. These are enriched in
the spilite fluid compared to the vent fluids. Accordingly,
the composition of the spilite fluid fits qualitatively well
with known chemical trends associated with the chemical
reequilibration of seawater as it moves through hot basal-
tic crust. Moreover, the analyzed spilite fluid, which
30 L. Richter, L.W. Diamond / Geochimica et Cosmochimica Acta xxx (2021) xxx–xxx
Please cite this article in press as: Richter L. and Diamond L. W. Characterization of hydrothermal fluids that alter the upper oceanic
crust to spilite and epidosite: Fluid inclusion evidence from the Semail (Oman) and Troodos (Cyprus) ophiolites. Geochim. Cosmochim.
Acta (2021), https://doi.org/10.1016/j.gca.2021.11.012
<