of 20
Chemical Geology 578 (2021) 120304
Available online 5 May 2021
0009-2541/© 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (
http://creativecommons.org/licenses/by/4.0/
).
Invited
Research
Article
Cold-water
corals
as archives
of seawater
Zn and Cu isotopes
Susan
H. Little
a
,
b
,
*
, David
J. Wilson
a
, Mark Rehk
̈
amper
b
, Jess F. Adkins
c
, Laura
F. Robinson
d
,
Tina van de Flierdt
b
a
Department
of Earth
Sciences,
University
College
London,
Gower
Place,
London
WC1E
6BS,
UK
b
Department
of Earth
Science
and Engineering,
Royal
School
of Mines,
Imperial
College
London,
London
SW7
2BP,
UK
c
Division
of Geological
and Planetary
Sciences,
California
Institute
of Technology,
Pasadena,
CA 91125,
USA
d
School
of Earth
Sciences,
University
of Bristol,
Bristol
BS8 1RJ,
UK
ARTICLE
INFO
Editor:
Michael
E. Boettcher
Keywords:
Cold-water
coral
Zinc
Copper
Isotopes
Aragonites
Palaeoceanography
ABSTRACT
Traditional
carbonate
sedimentary
archives
have proven
challenging
to exploit
for Zn and Cu isotopes,
due to the
high concentrations
of trace metals
in potential
contaminants
(e.g., Fe-Mn
coatings)
and their low concentrations
in carbonate.
Here, we present
the first dataset
of
δ
66
Zn
JMC-Lyon
and
δ
65
Cu
SRM 976
values
for cold-water
corals
and
address
their potential
as a seawater
archive.
Extensive
cleaning
experiments
carried
out on two corals
with well-
developed
Fe-Mn
rich coatings
demonstrate
that thorough
physical
and chemical
cleaning
can effectively
remove
detrital
and authigenic
contaminants.
Next,
we present
metal/Ca
ratios
and
δ
66
Zn and
δ
65
Cu values
for a
geographically
diverse
sample
set of Holocene
age cold-water
corals.
Comparing
cold-water
coral
δ
66
Zn values
to
estimated
ambient
seawater
δ
66
Zn values
(where
Δ
66
Zn
coral-sw
=
δ
66
Zn
coral
– δ
66
Zn
seawater
), we find
Δ
66
Zn
coral-sw
= +
0.03
±
0.17
(1SD,
n
=
20). Hence,
to a first order,
cold-water
corals
record
seawater
Zn isotope
com
-
positions
without
fractionation.
The average
Holocene
coral Cu isotope
composition
is
+
0.59
±
0.23
(1SD,
n
=
15), similar
to the mean
of published
deep seawater
δ
65
Cu values
at
+
0.66
±
0.09
, but with considerable
variability.
Finally,
δ
66
Zn and
δ
65
Cu data are presented
for a small subset
of four glacial-age
corals.
These
values
overlap
with the respective
Holocene
coral datasets,
hinting
at limited
glacial-interglacial
changes
in oceanic
Zn
and Cu cycling.
1. Introduction
Zinc (Zn) and copper
(Cu) are bioessential
trace metals
with isotopic
systems
that are emerging
as promising
tracers
of past ocean
nutrient
and redox cycling.
To date, reliable
archives
for past seawater
Zn and Cu
isotopes
have been lacking,
because
both metals
are present
at low
concentrations
in carbonate
and opal, but at high concentrations
in
potential
contaminating
material,
such as detrital
or authigenic
(e.g., Fe-
Mn oxide)
phases
(Boyle,
1981; Shen and Boyle,
1988; Pichat
et al.,
2003; Andersen
et al., 2011; Hendry
and Andersen,
2013). Zinc isotopes
have previously
been applied
in Pleistocene
to ancient
marine
sedi
-
ments,
typically
using bulk carbonate
leachates
in an attempt
to side-
step the contamination
problem
(Pichat
et al., 2003; Kunzmann
et al.,
2013; John et al., 2017; Liu et al., 2017; Sweere
et al., 2018). We pro
-
pose that cold-water
coral skeletons
provide
an exciting
new possibility
for a seawater
Zn and Cu isotope
archive.
Their
global
distribution,
combined
with an ability
to obtain
precise
ages for individual
specimens,
gives
corals
distinct
advantages
over more
traditional
palaeoclimate
archives
(e.g., Robinson
et al., 2014), potentially
enabling
reconstructions
of ocean
chemistry
on centennial
or even shorter
time
-
scales
(e.g., Wilson
et al., 2014; Chen et al., 2015). In addition,
their
large size confers
a particular
advantage
for analysing
trace metal
iso
-
topes,
because
it should
enable
rigorous
cleaning
to remove
surficial
contaminant
phases,
while still providing
sufficient
quantities
of Zn and
Cu for isotope
analysis.
Zinc has a classic
nutrient-type
distribution
in the modern
ocean,
reflecting
a combination
of biological
cycling
and the physical
ocean
circulation
(Bruland,
1980; Vance
et al., 2017; Middag
et al., 2019).
Away from local sedimentary
sources
and hydrothermal
vents,
the deep
ocean
is isotopically
homogeneous,
at about
+
0.45
(
δ
66
Zn relative
to
JMC-Lyon),
while
the upper
ocean
exhibits
considerable
variability,
ranging
from
1.1 to
+
1.2
(Conway
and John,
2014; Zhao et al.,
2014; Samanta
et al., 2017; John et al., 2018; Wang
et al., 2018; Vance
et al., 2019; Liao et al., 2020; Sieber
et al., 2020). The origin
of this
* Corresponding
author
at: Department
of Earth Sciences,
University
College
London,
Gower
Place,
London
WC1E
6BS, UK.
E-mail
address:
susan.little@ucl.ac.uk
(S.H. Little).
Contents
lists available
at ScienceDirect
Chemical
Geology
journal
homepag
e:
www.else
vier.com/loc
ate/chem
geo
https://doi.org/10.1016/j.chemgeo.2021.120304
Received
30 November
2020;
Received
in revised
form 23 April 2021;
Accepted
2 May 2021
Chemical Geology 578 (2021) 120304
2
variability
remains
a subject
of debate.
Biological
uptake
in the Southern
Ocean
is associated
with no isotopic
fractionation
or with small
de
-
viations
towards
isotopically
heavy
Zn in the residual
dissolved
phase
(Zhao et al., 2014; Wang et al., 2018; Sieber
et al., 2020), consistent
with
evidence
for limited
isotopic
fractionation
on uptake
by diatoms
(John
et al., 2007; K
̈
obberich
and Vance,
2017). However,
marked
deviations
towards
isotopically
light sub-surface
Zn are observed
in the water
column
elsewhere
(e.g., Conway
and John,
2014). Explanations
pro
-
posed
for this phenomenon
include
shallow
remineralisation
of isoto
-
pically
light organic
matter
(Samanta
et al., 2017; Vance
et al., 2019),
removal
by scavenging
of isotopically
heavy
Zn (Conway
and John,
2014; John and Conway,
2014; Weber
et al., 2018; Liao et al., 2020),
and/or
inputs
from an isotopically
light external
Zn source
(Lemaitre
et al., 2020; Liao et al., 2020).
Dissolved
Cu concentrations
in the ocean
increase
approximately
linearly
with depth
(Boyle
et al., 1977; Bruland,
1980). This profile
shape
has been described
as
hybrid-type
and attributed
to a combi
-
nation
of biological
uptake
and scavenging
(Bruland
et al., 2013).
However,
the respective
roles of (a) reversible
scavenging
in the water
column
and (b) irreversible
scavenging
followed
by benthic
sedimentary
input remain
to be fully deconvolved
(Boyle
et al., 1977; Little et al.,
2013, 2018; Roshan
and Wu, 2015; Richon
and Tagliabue,
2019). A key
feature
of Cu (and, to a lesser
extent,
Zn) biogeochemistry
is its near
ubiquitous
complexation
by strong
organic
ligands;
in the ocean
>
99%
of dissolved
Cu is organically
complexed
(e.g., Coale and Bruland,
1988;
Moffett
and Dupont,
2007; Bruland
et al., 2013). The isotopic
compo
-
sition
of Cu in seawater
is more sparsely
documented
than that of Zn.
Existing
data also point to a homogeneous
deep ocean,
at about
+
0.66
(
δ
65
Cu relative
to NIST SRM 976), with deviations
towards
lighter
Cu
isotope
compositions
(of about
+
0.30
) in the surface
ocean
and along
some continental
margins,
which
appear
to be associated
with particu
-
late Cu input (Takano
et al., 2014; Thompson
and Ellwood,
2014; Little
et al., 2018; Baconnais
et al., 2019).
Cold water
corals
grow at shallow
to lower
bathyal
water
depths
(bathyal
zone:
1000
4000
m), and occasionally
in deeper
waters
(Roberts
et al., 2009), and therefore
offer potential
as an archive
of in
-
termediate
and deep ocean
δ
66
Zn and
δ
65
Cu values.
While
the upper
water column
cycling
of Zn and Cu isotopes
is complex
and incompletely
understood,
the relative
homogeneity
of modern
intermediate
and deep
ocean
Zn (and presumably,
albeit
to a lesser
extent,
Cu) isotope
com
-
positions
reflects
the first-order
role of water
masses
originating
from
the Southern
Ocean
in setting
global
oceanic
nutrient
distributions
(Sarmiento
et al., 2004; Vance
et al., 2017; de Souza
et al., 2018; Sieber
et al., 2020). Today,
the absence
of significant
biological
Zn isotope
fractionation
in the surface
Southern
Ocean
(Zhao et al., 2014; Wang
et al., 2018; Sieber
et al., 2020) leads to the formation
and advection
northwards
of intermediate
(i. e. Sub-Antarctic
Mode Water,
SAMW,
and
Antarctic
Intermediate
Water,
AAIW)
and deep (Antarctic
Bottom
Water,
AABW)
water
masses
with the same (or very similar)
isotopic
compositions
(Sieber
et al., 2020). However,
the physical,
biogeo
-
chemical,
and ecological
characteristics
of the Southern
Ocean
have
changed
through
time (Sigman
et al., 2010). For example,
alleviation
of
Fe limitation
in the past may have dramatically
affected
the nutrient
status
(and isotopic
composition)
of glacial
analogues
of AAIW
and
SAMW,
as proposed
for Si (e.g., Brzezinski
et al., 2002; Matsumoto
et al.,
2002). Therefore,
intermediate
and deep-water
Zn and Cu isotope
compositions
archived
in cold-water
corals
could be used to trace past
changes
in biological
utilization
in the Southern
Ocean,
with implica
-
tions for the global
ocean
carbon
cycle.
In this study
we evaluate
the potential
of cold-water
corals
as ar
-
chives
of seawater
Zn and Cu isotopes.
We present
a series
of physical
and chemical
cleaning
experiments,
followed
by
δ
66
Zn,
δ
65
Cu, and trace
element
data for a suite of modern
and late Holocene
(
<
1500 yr old)
cold-water
corals
from six oceanic
regions
spanning
the North
Atlantic
to the Tasman
Sea. Coral aragonite
δ
65
Cu values
are distributed
around
the modern
deep ocean
Cu isotope
composition,
but exhibit
significant
scatter.
The outlook
for Zn is more promising,
with reasonable
agree
-
ment between
coral aragonite
δ
66
Zn values
and measured
or best esti
-
mate modern
seawater
δ
66
Zn values.
A small
number
of older fossil
specimens,
dated
to the last glacial
period,
were also analysed.
These
corals
have Zn and Cu isotope
compositions
similar
to modern
seawater
values,
hinting
at the relative
constancy
of oceanic
Zn and Cu cycling
on
glacial-interglacial
timescales.
2. Samples
and analytical
methods
2.1.
Samples
The term cold-water
coral (or alternatively
deep-water
coral)
is used
here to refer to azooxanthellate
scleractinian
corals,
of which
~90%
live
in deep or cold water (Roberts
et al., 2009). Specimens
in this study are
solitary
aragonitic
corals
of the species
Desmophyllum
dianthus
and
genera
Caryophyllia
and
Dasmosmilia
. Taxonomic
classification
of sam
-
ples was carried
out in previous
studies
(listed
below).
Caryophyllia
is the
most diverse
genus
of cold-water
corals,
consisting
of at least 66 species
(Kitahara
et al., 2010a). Genetic
studies
have highlighted
the similarity
of extant
Caryophyllia
and
Dasmosmilia
genera
(Kitahara
et al., 2010b
).
Two
D. dianthus
specimens
of glacial
age, with well-developed
black
or brown
surface
coatings
(due to the presence
of Fe-Mn
oxide phases),
were selected
for cleaning
experiments
(described
in Section
2.2).
Twenty
modern
or late Holocene
(
<
1500 yr old) coral specimens
were
then sampled,
cleaned
following
the finalised
cleaning
procedure
(Sec
-
tion 2.2), and analysed
for trace element
concentrations
and Zn and Cu
isotopes
(Table 1A). Finally,
four corals
dated to the last glacial
period
(including
the two specimens
used in cleaning
experiments)
were
sampled
and analysed
for trace element
concentrations
and Zn and Cu
isotopes
(Table 1B).
Coral samples
from water depths
of 170
2260 m were selected
from
the following
locations
(Fig. 1; Table
1): south
of Iceland
(Reykjanes
Ridge),
the northwest
Atlantic
(Manning
and Muir Seamount),
the
eastern
equatorial
Atlantic
(Carter
Seamount),
the Drake
Passage
(Burdwood
Bank and Sars Seamount),
the southwest
Indian
Ocean
(SWIO:
Coral
Seamount,
Melville
Bank,
Atlantis
Bank),
and south
of
Tasmania
(South
Hills and St Helens
seamounts).
The corals
selected
from these collections
have previously
been described
and dated
by
uranium-series
or radiocarbon
in: Burke
(2012)
, Robinson
et al. (2007)
,
Chen et al. (2016)
, Margolin
et al. (2014)
, Pratt et al. (2019)
, and
Thiagarajan
et al. (2013)
.
2.2.
Cleaning
experiments
Fossil
cold-water
corals
are often coated
with a black-brown
crust,
made
up of a mixture
of iron and manganese
oxides,
incorporated
detrital
aluminosilicate
grains,
and occasional
metal
sulphides
(Cheng
et al., 2000). All these potential
contaminating
phases
contain
trace
metals
like Zn and Cu in concentrations
that are orders
of magnitude
higher
than those in coral aragonite
(e.g., Boyle,
1981; Shen and Boyle,
1988). We tested
the effectiveness
of the physical
and chemical
cleaning
procedures
developed
previously
for cold-water
corals
(Shen and Boyle,
1988; Lomitschka
and Mangini,
1999; Cheng
et al., 2000; van de Flierdt
et al., 2010; Crocket
et al., 2014) for the analysis
of Zn and Cu isotopes.
Two
D. dianthus
corals
with a well-developed
coating
were selected
for the cleaning
experiments:
DH115-DC-01
from the Drake
Passage,
and SS0108
from Tasmania
(Table
1B). The black coatings
of the two
specimens
were collected
using a scalpel
(coating
samples
were desig
-
nated ‘Coat
). Thereafter,
100
150 mg coral sub-samples
were obtained
using a Dremel
tool and progressively
subjected
to increasingly
rigorous
cleaning
steps (Fig. 2):
First, sub-samples
of ‘uncleaned
coral (designated
‘UNCL
) were
rinsed
three times in DI water,
where
rinsing
refers to ultrasonication
S.H. Little
et al.
Chemical Geology 578 (2021) 120304
3
Table
1
Location,
taxonomic
classification,
water
depth
and water
mass, and age of cold-water
coral specimens
included
in this study.
S.H. Little
et al.
Chemical Geology 578 (2021) 120304
4
of the coral fragments
in DI water
in acid-cleaned
15 mL centrifuge
tubes for 60 s, followed
by removal
of the supernatant
by pipette.
Second,
three to four sub-samples
were carefully
physically
cleaned
using a Dremel
tool and rinsed
three times in DI water
(designated
‘PHYS
). During
physical
cleaning,
any coatings
were removed,
as
well as epibiont
boreholes
and other
discoloration
or impurities
within
the skeleton,
including
remineralised
or secondary
calcium
carbonate.
Third,
three to four sub-samples
were subject
to physical
cleaning
and an oxidising
chemical
‘pre-cleaning
procedure
(designated
‘OXIC
). The chemical
pre-cleaning
procedure
consisted
of ultra
-
sonication
in a series
of oxidative
cleaning
solutions,
targeting
re
-
sidual
organic
phases
(Shen and Boyle,
1988), with rinses in DI water
(for details,
see Fig. 2).
Finally,
three to four sub-samples
were subjected
to the full physical
and chemical
cleaning
procedure
described
by van de Flierdt
et al.
(2010)
and detailed
in Fig. 2 (designated
‘FULL
). The major
differ
-
ence between
the pre-cleaning
and full chemical
cleaning
procedures
is the addition
of a reductive
cleaning
step, which
aims to remove
trace metals
associated
with residual
iron and manganese
oxides
(van de Flierdt
et al., 2010).
For all other coral specimens
for which
data are reported
in this
study,
sub-samples
of 80
200 mg were subjected
to the full physical
and
chemical
cleaning
procedure
(Fig. 2). Chemical
cleaning
led to an
average
mass loss of 7.0
±
3.5% (1SD,
n
=
36, range 1.9
21%).
Where
possible,
samples
were analysed
for Zn and Cu isotope
compositions
in
duplicate
(i. e. two coral sub-samples
were separately
cleaned,
digested,
and analysed).
Four of the Holocene
corals
and all four glacial-age
coral
specimens
had well-developed
coatings
(Table
1), which
were also
analysed
to evaluate
the possibility
of residual
contamination
by Fe
Mn
oxide or detrital
(aluminosilicate)
phases.
2.3.
Analytical
procedures
Sample
digestion
and column
chemistry
was carried
out in the
MAGIC
clean
laboratories
at Imperial
College
London.
Throughout,
deionized
18.2 M
Ω
water
(MQ),
Teflon-distilled
acids (HNO
3
and HCl),
Suprapur
H
2
O
2
, and acid-cleaned
Savillex
PFA labware
were used. In
preparation
for analysis,
corals
and coatings
were carefully
bulk diges
-
ted in 1 mL 6 M HCl (carbonate
effervesces
vigorously
on addition
of
acid).
For coral samples
with significant
coatings,
residual
contamina
-
tion was assessed
using
a mass balance
approach
and found
to be
negligible
(see Section
4.2, Table
S3). Samples
were then dried and
redissolved
in 5 mL 1 M HCl.
An aliquot
of this coral digest
solution
was diluted
in 2% HNO
3
to
give approximately
100 ppb Ca concentrations
for multi-element
anal
-
ysis on a Thermo
Element
XR at ETH Zürich.
The elements
Li, Na, Mg, Al,
Mn, Fe, Cu, Zn, Sr, Cd, Ba, and U were measured
as metal/Ca
ratios
following
the procedure
outlined
in Hasenfratz
et al. (2017)
. Accuracy
and precision
of the instrument
were assessed
by routine
measurements
of four consistency
standards
(Table
S1), of which
three are gravimet
-
rically
prepared
in-house
standards
(CS1, CS2, CS3) and one is a car
-
bonate
rock standard
purchased
from LGC Standards
(NCS DC70303).
The appropriate
volume
of a
64
Zn-
67
Zn double
spike was added
to all
samples
to achieve
a spike:sample
ratio of approximately
1.2 (Arnold
et al., 2010; Bridgestock
et al., 2014). In order to obtain
40 ng Zn for
isotope
analysis,
some sub-samples
from the cleaning
experiments
were
combined
(Table 2). The Zn and Cu fractions
were then purified
using a
two-step
column
chromatography
procedure
using
AG MP-1 M resin
(BioRad),
as detailed
previously
(Mar
́
echal
et al., 1999; Archer
and
Vance,
2004; Little et al., 2014). The second
Zn column
was smaller
in
volume,
following
Bridgestock
et al. (2014)
. Prior to analysis,
purified
Zn and Cu fractions
were oxidised
by treatment
with 2
×
~100
μ
L 14 M
HNO
3
(Zn) or refluxing
overnight
with 14 M HNO
3
+
H
2
O
2
(Cu), before
final dissolution
in 2% HNO
3
. Procedural
Zn and Cu blanks
were
<
1 ng.
References:
Burke
(2012)
: Iceland:
14
C.
Margolin
et al. (2014)
: Drake
Passage:
U-series.
Chen et al. (2016)
: Equatorial
Atlantic:
U-series.
Pratt et al. (2019)
: Southwest
Indian
Ocean:
U-series.
Robinson
et al. (2007)
: Northwest
Atlantic:
U-series.
Thiagarajan
et al. (2013)
: Tasman
Sea:
14
C.
DC and SS: Cleaning
specimens;
see Table 2.
Lat/Long
in italics:
approximate
location.
S.H. Little
et al.