Chemosphere 352 (2024) 141487

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Practical guidelines for representing and interpreting rare earth

abundances in environmental and biological studies
Jean-Alix Barrat a, b, *, Germain Bayon c
a
Univ Brest, CNRS, Ifremer, IRD, LEMAR, Institut Universitaire Européen de la Mer (IUEM), Place Nicolas Copernic, 29280, Plouzané, France
b
Institut Universitaire de France, Paris, France
c
Univ Brest, CNRS, Ifremer, Geo-Ocean, F-29280, Plouzané, France

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Rationale behind the use of REE

normalization.
• Misuse and common errors in drawing
REE patterns.
• REE patterns of living organisms show
promising perspectives.

A R T I C L E I N F O A B S T R A C T

Handling editor: Milena Horvat This paper summarizes the main guidelines for representing rare earth element (REE) abundance patterns, along
with a review of the common mistakes or omissions that can alter REE plots and bias interpretations. It is
Keywords: specifically designed for ecotoxicologists and biologists, for whom the study of these elements has become an
Rare earth elements important field of research in recent years. Prior to applying REE diagrams to the study of living organisms, it is
Oddo-harkins effect
important to understand the rationale that led geochemists and cosmochemists to develop them. Used with the
REE pattern
practical recommendations described here, these diagrams have the capacity to highlight fundamental processes
Normalization values
REE anomalies taking place in the biosphere.

* Corresponding author. Univ Brest, CNRS, Ifremer, IRD, LEMAR, Institut Universitaire Européen de la Mer (IUEM), Place Nicolas Copernic, 29280 Plouzané,
France.
E-mail address: [email protected] (J.-A. Barrat).

https://doi.org/10.1016/j.chemosphere.2024.141487
Received 27 November 2023; Received in revised form 20 January 2024; Accepted 15 February 2024
Available online 17 February 2024
0045-6535/© 2024 Elsevier Ltd. All rights reserved.
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

1. Introduction 2. A few words on analytical techniques

Rare earth elements (REEs) are a group of fifteen elements - the Over the past 60 years, considerable progress has been made in
lanthanides (La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu), determining REE abundances in rocks and water. Most analytical tech­
with Z values ranging from 57 to 71, to which chemists commonly add Y niques have been applied with varying degrees of success, depending on
(Z = 39) and sometimes Sc (Z = 21), although the latter has a very the type of sample. Until the 1990s, these techniques were not routinely
different geochemical behavior. Apart from Pm, which is radioactive, capable of determining all REEs. They also required complex equipment,
with a very short half-life (only 2.62 y), and is therefore absent in nature, and were frequently time-consuming and laborious. For example,
these elements are not so rare and typically occur in trace amounts in neutron activation, which was a major analytical technique, required
most rocks. Over the past 60 years, REEs have become one of the most the use of a nuclear reactor (neutron source) to irradiate samples, and
studied groups of elements in Earth and Universe Sciences. It is beyond the use of scintillation-type detectors to record the spectra of radioactive
the scope of this paper to enumerate all the diverse applications offered emissions to calculate element concentrations at different times after
by REEs. Just as an example, the behavior of the REEs during magma irradiation. Several counting stages, spaced several tens of days apart,
genesis is well understood, so their abundances can be used in petro­ and sometimes several months, were required to obtain La, Ce, Nd, Sm,
logical studies to provide constraints on the formation of both terrestrial Eu Tb, Yb and Lu (e.g., Chayla et al., 1973). At the time, the most ac­
and extraterrestrial rocks. Additionally, REE are powerful proxies to curate measurement technique was isotope dilution combined to ther­
investigate both present and past surface processes, in both marine and mal ionization mass spectrometry (ID-TIMS; e.g., Schnetzler and
continental environments, for characterizing and tracing ocean water Philpotts, 1968; Gast et al., 1970). Samples in solution were spiked with
masses (e.g., Garcia-Solsona et al., 2014), but also estuarine processes (e. a solution containing artificial REEs whose compositions were very
g., Elderfield et al., 1990), groundwater mixing (e.g., Johannesson et al., different from those of natural REEs. The REEs in the spiked sample were
1997; Noack et al., 2014), and for understanding carbon sequestration separated using ion exchange columns, then their isotopic compositions
(e.g., Karamalidis et al., 2013), acid mine drainage (Prudêncio et al., were measured using solid-source mass spectrometry. This technique
2015), and in-situ remediation of groundwater (e.g., Wilkin et al., solely allowed measurement of the concentrations of polyisotopic rare
2021). Furthermore, the radiogenic isotopic compositions of REE (i.e. earths (La, Ce, Nd, Sm, Eu, Gd, Dy, Er, Yb, Lu), albeit with very high
Nd, Ce) are now well-established tools for dating rock formation, in accuracy (RSD<2–3%). It was nevertheless time-consuming, since one
provenance studies, and to reconstruct the evolution of geochemical day of measurement was required to determine the REE concentrations
reservoirs during our planet’s history (e.g., Henderson, 1984). of any single sample by single-collector mass spectrometry at that time.
The tremendous interest drawn for rare earths by cosmochemists, Additionally, prior to analysis, the time needed to prepare a series of
geochemists and geologists over the last decades is in stark contrast to samples could approach 3 weeks, including rock dissolution, beaker
the much more recent attention paid to these elements in life sciences. In washing and decontamination, and chemical separations.
contrast to most geological samples, REE abundances in organisms are Other techniques such as inductively coupled plasma - atomic
generally very low. It is only very recently that these elements have been emission spectrometry (ICP-AES) were subsequently developed,
shown to participate in biochemical cycles, particularly in methano­ enabling REEs to be determined simultaneously and more rapidly (e.g.,
trophic organisms (e.g., Pol et al., 2014; Semrau et al., 2018). Apart from Walsh et al., 1981; Watkins and Nolan, 1992). While this technique is
these recent discoveries, the potential use of REEs by organisms remains still used by some laboratories, ICP-AES instruments were superseded in
largely unknown. While evidence exist for REE accumulation in certain the late 1990s by the development of inductively coupled plasma-mass
organisms (e.g., Ozaki et al., 1997) and for their utility in promoting spectrometry (ICP-MS), enabling all REEs to be determined, with very
plant growth (e.g., Ouyang et al., 2003), the mechanisms involved in the low detection limits (e.g., Jenner et al., 1990; Jarvis and Jarvis, 1992)
biological use of REE are still poorly understood. Substantial research and generally much improved accuracy for low-level samples charac­
effort will be required in future years on the above-mentioned subjects, terized by low REE concentrations.
especially as these elements are nowadays brought into the environment With the advent of ICP-MS, the acquisition of high-quality data with
by human activities. Since the 1970s, the industrial demand for REEs has accuracies equivalent to those yielded by isotopic dilution techniques in
grown exponentially. These metals have become highly strategic (e.g., the 1990s has now become routinely accessible and fast, with analytical
Hatch, 2012; Massari and Ruberti, 2013) and are now indispensable in costs being much lower than with neutron activation or ID-TIMS.
many critical fields, such as electronics, manufacturing of magnets (e.g., Nowadays, the time required to analyze a sample for REE concentra­
for wind turbines), batteries, oil refining (fluid-cracking catalysts), tions by ICP-MS only takes a few minutes (excluding sample preparation
polishing medias, and medical imaging. A corollary of this markedly time). Almost all REE data are now acquired by ICP-MS.
increasing consumption of REE is the emergence of new pollution In contrast to most geological materials, biological samples are
pathways linked to the extraction, purification and use of these ele­ commonly depleted in REEs, with concentrations for these elements
ments, and above all to the production of new waste (e.g., Brewer et al., typically in the range of a few ng/g to tens of ng/g for some of them (e.g.,
2022). This raises new questions for the environmental sciences and Danezis et al., 2019 and references therein). As a consequence, the
biology: how can we identify and quantify the pollution generated by analysis of biological samples for REE is even more difficult than for
these elements? What are the impacts of these emerging pollutants on rocks displaying the lowest REE abundances. Regarding the analytical
the environment and on living organisms? To answer these questions, measurement of REE by ICP-MS, two important points have to be
but also to understand how these elements may be transferred from emphasized:
nutrients to organisms, from one organ to another, but also on another -First, particular care needs to be taken regarding measurement
scale from one trophic level to another, diagrams need to be developed quality and instrumental conditions, as isobaric interferences can be
by biologists and ecotoxicologists. Such diagrams, known as generated in the plasma. These interferences are capable of generating
REE-patterns, are widely used in geochemistry and cosmochemistry. Not significant concentration anomalies that can result in biased measured
only do they enable REE abundances to be visualized, they can also be abundances. As for the geological samples, REE interferences should be
used to detect enrichment or depletion of certain REEs relative to systematically monitored and corrected during every ICP-MS analytical
neighboring elements and corresponding so-called ‘anomalies’. The aim session. Eu concentrations cannot be determined correctly if isobaric Ba
of this paper is to introduce the basic principles of these diagrams first oxide interferences are not properly estimated and corrected. This type
developed for applications in Earth Sciences, but showing great poten­ of artifact is known for plants (Pourret et al., 2022), and is particularly
tial in the fields of biochemistry and ecotoxicology. critical for samples displaying high Ba/REE ratios. It can be easily
detected using Eu/Sm vs. Ba/Sm diagrams (Fig. 1). The reader is

2
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

biological studies. Moreover, organic matter can be depleted in REEs by

several orders of magnitude compared to associated detrital material (e.
g., soil, sediment, dust, oxides). As a result, some biological standards
may contain minute amounts of inorganic (terrigenous) matter, much
enriched in REEs, which can affect measured REE abundances depend­
ing on the procedures used for sample preparation and digestion. The
“duckweed” reference standard (BCR 670) is a good example of this
problem (Zocher et al., 2021). The preparation of a rock powder is of
course very different from that of organic matter, but it will be important
for biological studies in the future to have access to well-characterized
‘organic’ reference materials for REE abundances, as well some of the
precisely characterized geostandards (e.g., BCR-2 or BHVO-2). This
would validate observed anomalies and facilitate intercalibration of
results between different laboratories worldwide. Biologists must bear in
mind that, behind the questions of bioaccumulation and vital effects, the
REEs contained in well-prepared samples of organisms originate from
their living environments, and that consequently, the distributions
observed in organic matter are largely inherited from those found in
ambient water and sediments. Therefore, any REE ‘anomaly’ identified
in biological samples can only be directly attributed to vital effects or
anthropogenic contributions providing that interferences and calibra­
tion have been properly assessed during analysis.

3. Why is the behavior of REEs so coherent and why do some

REEs decouple from the rest of the group?
Fig. 1. Eu/Sm vs. Ba/Sm plot for waters from St. Lawrence River (Dang et al.,
2022). Measurements were obtained by ICP-MS. Eu/Sm ratios correlate with REEs are remarkable in many respects. First, their atoms have
Ba/Sm ratios (r = 0.95). Such a relationship indicates that Eu concentrations complex electronic configurations, with a 4f electronic sublayer filling in
are affected by isobaric BaO interferences generated in the plasma, and are
from Ce to Lu. Under terrestrial conditions, these elements are generally
mostly incorrect. Dang et al. (2022) also analyzed a control standard with
trivalent. The consequence of these electronic configurations is that
satisfactory results. This standard (SLR6) has a Ba/Sm ratio of around 400, too
low to detect the effect of interferences. their ionic radius decreases linearly with the atomic number (Fig. 2).
This contraction in ionic radius is known to chemists as “lanthanide
contraction”. From La3+ to Lu3+, effective ionic radii (coordination
referred to Zepeda et al. (2023) for a recent discussion.
number VIII) change from 1.16 (La) to 0.977 Å (Lu), a variation of only
In samples showing extremely high Ba/REE ratios, isobaric in­
18.7% (Shannon, 1976). Between two consecutive elements, effective
terferences can even affect Sm and possibly Gd abundances. Apart from
ionic radii vary by only 0.8–1.5 %. As the chemical behavior of ions is
these particular samples, interference corrections are generally suc­
largely controlled by charge (3+ here), and ionic radius, these variations
cessful for determining reliable and accurate Eu abundances. Alterna­
explain on the one hand the consistency of the chemical behavior of
tively, the use of ICP-MS instruments equipped with collision cells that
trivalent REEs in nature, but also the impossibility for natural processes
can efficiently break up complex molecules formed in the plasma can
to specifically decouple one of these elements from its neighbors.
quantitatively reduce isobaric interferences. Additionally, high-
Consequently, the abundances of two neighboring trivalent REEs (e.g.,
resolution sector field ICP-MS offers the capability of resolving REE
Pr and Nd, Gd and Tb, Tb and Dy, Dy and Ho, Ho and Er, etc.) are
ion signals from their isobaric interferences. Despite these advantages,
these techniques cannot be applied to REE-poor samples, since they
typically result in a drastic decrease of signal intensity by two orders of
magnitude. Moreover, they do not generally produce better quality re­
sults than those obtained at low resolution with interference correction
for regular samples (Charles et al., 2021). Therefore, for Ba-rich samples,
the best solution to overcome interference issues is to separate Ba from
the REEs prior to analysis. Several methods have been proposed and
successfully applied routinely for REE-depleted samples (e.g., Barrat
et al., 1996, 2020; Bau et al., 2010).
Similarly, the oxides and hydroxides of light REEs and middle REEs
formed in the plasma must be evaluated and accounted for during
analytical sessions, as they correspond to isobars of the isotopes moni­
tored to determine Gd and heavy-REE concentrations, hence possibly
leading to overestimation of their measured abundances. This problem
is particularly critical in the case of Gd pollution studies (see below).
-Second, a large number of certified reference materials for geolog­
ical materials now exist, for which very precise and accurate REE con­
centrations have been determined. Among these, two basaltic materials -
BHVO-2 (Hawaii) and BCR-2 (Colombia River) - are among the best
characterized geological standards (Jochum et al., 2016) available at
present. Unfortunately, there are currently no biological standards as
precisely characterized for REE abundances. This lack of standard rep­
resents one major limitation for the validation of REE determinations in Fig. 2. Effective ionic radii vs. atomic numbers (Z) for REE in VIII coordination
(Shannon, 1976).

3
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

generally highly correlated in nature, and the linear relationships ob­ of REEs has only been considered recently. Biological effects on REE
tained between these pairs of elements are therefore useless, and pro­ abundances can be hampered by environmental signals. For instance,
vide no constraint for discussing/discriminating the processes that thiotrophic mussels living in the vicinity of submarine hydrothermal
might be responsible for variations in abundance as exemplified in vents typically show significant excess in Eu abundances, which reflect
Fig. 3. the composition of hydrothermal fluids (Bau et al., 2010; Barrat et al.,
Natural abiotic processes can modify the relative proportions of REE 2022a), not biological effects. In the case of Ce, fractionation linked to
abundances. For example, the partial melting of rocks in the Earth’s biological activity has been recently demonstrated by Kraemer and Bau
mantle will produce lavas enriched in REE compared to corresponding (2022). Filter-feeding bivalves (dog cockles, Glycymeris) also indicate
parental rocks, generating enrichments in the lightest REEs. These en­ significant Ce fractionation that co-varies with the age of the animals,
richments will not be specific to any single element, but progressive and hence suggesting a possible metabolic link (Barrat et al., 2022b). These
a function of the element’s atomic number (Z) or ionic radius. Such studies are promising and call for further in-depth investigation of the
gradual enrichments can be modeled in order to infer magma formation mechanisms driving biological fractionation in living organisms.
models for example (e.g., Minster and Allègre, 1978). However, specific Lanthanum excess is typically observed in seawater and in various
decoupling of particular elements from the other REEs is naturally chemical precipitates, in which this particular characteristic can be used
possible on Earth, essentially for the following elements: La, Ce, Eu and as a diagnostic feature of a seawater origin (e.g., Bau and Dulski, 1996a;
Y. Kamber and Webb, 2001). Lanthanum has been suggested to be more
Europium and Ce can have different valences, and therefore exhibit stable than other light REEs during complexation in seawater (de Baar
distinct behavior in Earth’s systems. During magma genesis, depending et al., 1985) or to be preferentially released from suspended barite
on oxygen fugacity, Eu can be present in two valence states, Eu2+ and particles (Grenier et al., 2018). None of these explanations is fully
Eu3+ (e.g., Goldschmidt, 1958). Europium2+ has an ionic radius similar satisfactory. Recently, it has been discovered that La is used by meth­
to that of Ca2+, and can therefore be substituted for this element in the anotrophic organisms (e.g., Pol et al., 2014; Semrau et al., 2018). These
crystal lattice of calcic minerals. Cerium is always trivalent during microbes convert methane into methanol, then into formaldehyde, using
magma genesis, but in aquatic environments, it can be oxidized and methanol dehydrogenase enzymes. Some methanotrophic organisms use
become tetravalent (e.g., Goldberg, 1961). This property explains the lanthanide-dependent enzymes to produce formaldehyde. This enzy­
distinctive behavior of Ce in oceans (e.g., German and Elderfield, 1990). matic activity is capable of generating La anomalies and other light-REE
This decoupling of Ce from the other REEs is an extremely useful tool for enrichments in marine organisms (Wang et al., 2020; Bayon et al., 2020;
reconstructing redox environmental changes over the Earth’s history (e. Barrat et al., 2022a), which can potentially impact REE abundances in
g., Tostevin et al., 2016; Wallace et al., 2017; Bellefroid et al., 2018 and surrounding water (Shiller et al., 2017). Recent experimental work has
references therein). also demonstrated the activity of Eu as an enzymatic cofactor in the
In life sciences, the possibility of significant biological fractionation methanotrophic bacterium Methylacidiphilum fumariolicum (Jahn et al.,
2018). Lanthanide-dependent methanotrophy is currently the only
natural process known to significantly fractionate REEs from each other,
but its potential importance in the marine REE budget remains to be
investigated.
As for REEs, Y (which is not a lanthanide) is also trivalent in nature
and has the same ionic radius as Ho3+. Unsurprisingly, the behavior of
these two elements is identical in a wide range of environments. In fact,
these elements are generally considered as geochemical twins. The Y/Ho
ratios in most chondrites, achondrites, Martian, lunar and terrestrial
rocks (magmatic and terrigenous sediments) vary within a very limited
range of values (~27.7 ± 2.7; Bau, 1996; Pack et al., 2007). However, in
aquatic environments, Y is generally strongly decoupled from Ho partly
due to different complexation behavior onto suspended particulates (e.
g., Bau et al., 1997; Nozaki et al., 1997). As a result, measured Y/Ho
ratios for seawater and seawater-derived precipitates (carbonates,
Fe–Mn crusts) significantly depart from chondritic or solid-Earth values,
meaning that Y/Ho can be used as a tool to discriminate the relative
contribution of marine versus terrestrial REE signatures in environ­
mental and biological samples (e.g., Kamber and Webb, 2001).

4. REE and REY patterns

In the universe, elements with an even Z number greater than 6 (i.e.

carbon) have higher abundances than their immediate neighbors with
an odd Z value. This property, known as the Oddo-Harkins effect, is a
consequence of nucleosynthesis, the origin of chemical elements. Rare
earths are no exception to this rule, and their relative abundances in
Fig. 3. Er vs. Ho plot for Martian rocks, terrestrial basalts, bivalves (soft tissues nature are largely inherited from primordial distributions. Geochemists
and shells), CI chondrite and post-archean shale average (PAAS). All the data realized early on that, in order to interpret and visualize REE abun­
were produced in the same laboratory and selected in order to avoid any
dances in samples, it was necessary to find a way of representing the data
analytical bias (Barrat et al., 2012, 2016, 2022b) except the shale average
that would erase the Oddo-Harkins effect and enable enrichments or
(Pourmand et al., 2011). All analyses lie on the same line, controlled by the
chondritic (CI) Er/Ho ratio. A wide variety of processes produced the observed depletions of REEs to be highlighted in relation to a reference. We owe
abundance ranges, but did not decouple these two elements. Obtaining a the creation of REE patterns to Masuda (1962) and Coryell et al. (1963).
straight line in this diagram does not discriminate between different processes Their principle is very simple: REE concentrations in a sample are
or sample origins, since samples that have absolutely nothing to do with each normalized to the abundances of a chosen reference material (Table 1).
other are always on this line. The resulting ratios or “normalized concentrations” are plotted from

4
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

Table 1 extraterrestrial or terrestrial materials, including biological objects.

Preferred normalization values. The chondrite (CI) is obtained from 5 g of the Second, this representation can be used to highlight “enrichments” or
Orgueil meteorite (Barrat et al., 2012). This value is preferred to other compi­ “depletions” in certain groups of REEs, commonly referred to as light
lations due to the large mass used. The sediment average (PAAS) is that obtained REEs (from La to Sm, “LREE”), middle REEs (Sm, Eu and Gd, “MREE”),
by Pourmand et al. (2011) and has been adjusted to the values of the standards and the heavy REEs (from Gd to Lu, “HREE”). Let us pause for one
obtained in Brest. Note that the Y/Ho ratio of the CI chondrite is 27.5, and that
moment to clarify what these words mean, which have nothing to do
the calibration we use is in agreement with the values of the standards given by
with the level of REE abundances in relation to the reference. Whatever
Jochum et al. (2016). The concentrations are given in μg/g and μmol/kg.
the REE abundances of any sample are, this particular sample will be
CI-chondrite CI-chondrite PAAS PAAS
considered as being enriched in light REEs if normalized abundances
unit μg/g μmol/kg μg/g μmol/kg decrease from La to Sm. This enrichment can be quantified with the Lan/
Smn ratio (also noted as (La/Sm)n; with n stating for ‘normalized’),
Y 1.56 17.55 32.2 362 which in this case is > 1. Conversely, a sample will be considered as
La 0.235 1.692 44.75 322.2 being depleted in light REEs when its normalized abundances increase
Ce 0.600 4.28 87.29 623
from La to Sm, hence characterized by (La/Sm)n < 1. The same defini­
Pr 0.091 0.646 10.1 71.68
Nd 0.464 3.22 36.98 256.4 tions equally apply to heavy REE; a sample enriched in heavy REE will
Sm 0.153 1.018 6.908 45.94 display a (Lu/Gd)n ratio greater than 1, and vice versa. In Fig. 3, we have
Eu 0.0586 0.386 1.188 7.818 plotted some patterns of typical rocks. The mid ocean ridge basalt
Gd 0.206 1.31 5.958 37.89 (MORB) is depleted in light REE compared to the reference chondrite,
Tb 0.0375 0.236 0.894 5.625
even though its REE abundances (including light REEs) are much higher
Dy 0.254 1.563 5.272 32.44
Ho 0.0566 0.343 1.078 6.536 than those of chondrites. The average shale (Post Archean Australian
Er 0.166 0.992 3.094 18.50 Shale or PAAS; Taylor and McLennan, 1985) is enriched in light REE but
Tm 0.0262 0.155 0.468 2.77 not in heavy REEs.
Yb 0.168 0.971 3.028 17.5
The resulting REE patterns can be used to highlight the decoupling of
Lu 0.0246 0.141 0.438 2.503
certain REEs such as La, Ce and Eu from neighboring REE, the calcula­
tion of which will be explained below. It should be emphasized here that
La to Lu following increasing atomic number Z (Fig. 4). As the reference raw (unnormalized) REE abundance patterns do not allow to read all the
also has elemental distributions reflecting the Oddo-Harkins effect, the information they carry. Therefore, these plots are obsolete and should
latter disappears, and a smooth curve is obtained that enables small not be used anymore. When describing REE patterns normalized to a set
differences in the fractionation of one REE from another to be identified of reference values, we are of course interested in the level of concen­
graphically. Pioneering authors of this representation chose a reference trations relative to this reference. We also attach great importance to the
chondritic material for normalization. general shape of the patterns, which may display any depletion,
There were several reasons for this choice. First, chondrites are enrichment, concavity or convexity, relative to the reference material
primitive meteorites, whose REE concentrations are proportional to used for comparison. These REE characteristics can be quantified, and
those of the solar photosphere, hence considered to represent the bulk along with concentration levels, are parameters that can be used to test
Solar System for these elements, and the building blocks of planets. This models concerning rock formation, or the origin of REEs accumulated in
reference is therefore universal, and can be used for all types of organisms.

Fig. 4. a/Unnormalized REE pattern (a) and chondrite normalized REE pattern for the reference CI chondrite, a mid-ocean ridge basalt (MORB), the reference post-
archean Australian shale average (PAAS), a Mn-nodule, and a duckweed sample. The x-axis corresponds to the atomic number of the elements (Z), which are linked to
the ionic radii (the ionic radius increases with decreasing Z, see Fig. 2). The non-normalized patterns have a sawtooth shape characteristic of the Oddo-Harkins effect.
They are uninterpretable, as enrichments and anomalies cannot be identified. Note that Pm is lacking in natural samples and cannot be placed in Fig. 4a. REE patterns
display a segment connecting Nd and Sm normalized concentrations by convention (Fig. 4b).

5
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

Geochemists working on sediments, waters and, more generally, and therefore cannot be measured. It is consequently absent from the list
surface processes, prefer to use a reference other than chondrites to of elements supplied by analysts, and is therefore commonly forgotten
normalize their data. They generally use average values for post- when REE patterns are drawn. The omission of Pm in normalized REE
Archean terrigenous sediments such as shales or muds, hence with patterns can lead to errors when calculating anomalies (see below), and
geologic age <2.5 billion years old (this is because the oldest Archean hence should be avoided.
continental crust was significantly different in composition; e.g. Taylor Yttrium can also be added to the elements to be represented in
and McLennan, 1985). Over the past decades, different set of values for normalized REE patterns. As discussed above, Y has a similar valence
terrigenous sediments have been proposed and used for REE normali­ and ionic radius as Ho, and hence can be inserted between Dy and Ho.
zation purposes, which are known by the following acronyms: PAAS This addition is likely to modify the smoothness in the heavy-REEs
(Nance and Taylor, 1976; Taylor and McLennan, 1985), NASC (Gromet portions of the patterns. However, the advantage of being able to
et al., 1984), MUQ (Kamber et al., 2005), WRAS and WRAC (Bayon highlight Y–Ho decoupling, particularly in aquatic environments,
et al., 2015), EUS (Bau et al., 2018). All these references serve the same largely offsets this drawback. These patterns are known as Rare Earth
purpose. Their REE abundances may slightly vary from one set of values and Y (REY) patterns (e.g., Bau and Dulski, 1996a). We would like to
to another, but they are generally taken as being representative of the draw readers’ attention to a calibration problem. While there is good
general REE composition of the upper continental crust, hence dis­ agreement on standard REE concentrations, state-of-the-art data diverge
playing concentration ratios presumably very close to those for this slightly on Y abundances. There is a small analytical bias between lab­
geochemical reservoir. Standardizing data with such a reference is oratories for this element, which can be higher than 10 %. This is
particularly recommended when investigating the geochemistry of illustrated by the average Y/Ho ratios (concentrations in μg/g) obtained
sediments or sedimentary rocks (including marine precipitates) and by various laboratories for international basalt and granite standards:
natural waters (river and marine waters), because their REE abundances 27.73 (σ = 1.16, n = 19), Jochum et al. (2016); 25.34 (σ = 0.55, n = 6),
are directly inherited from the upper continental crust. Therefore, the Pourmand et al. (2011); 23.63 (σ = 0.61, n = 8), Makishima and
shape of the REE pattern obtained with this type of reference material Nakamura (2006). We therefore recommend to check the standard re­
makes it easier to visualize elemental decoupling during Earth surface sults and choose normalization values that are compatible with the
processes or chemical precipitation in natural waters. Importantly, these calibration of the results that the reader wishes to normalize.
sedimentary references do not display any significant anomaly in La or In principle, there are two types of vertical scale (y-axis) that can be
Ce compared to chondrites, and consequently do not result in particular used for representing normalized REE patterns: a linear scale or a log­
bias when calculating corresponding anomalies for these elements, arithmic scale. The use of a logarithmic scale is often justified by the fact
whose values obtained are similar to those calculated with chondrites that normalized concentrations may vary over several orders of
(Barrat et al., 2023). Like the upper continental crust, these sediments magnitude. This is true, but not the only essential reason. Unlike linear
are enriched in light REE and have a negative Eu anomaly relative to scales, logarithmic scales preserve elemental ratios independently of
chondrites. Therefore, REE normalization using sediment reference their abundances. To illustrate this, rather than to discuss it mathe­
values offers the advantage of improving the visualization of REE pat­ matically, we have chosen to use a simple example. In Fig. 5, we have
terns for soils, continental sediments, marine sedimentary rocks and plotted the REE pattern of a seawater sample, and theoretically calcu­
natural waters, but they may also blur particular Eu anomalies inherited lated the patterns of the same seawater diluted with pure water. Such
from crustal reservoirs. Europium is trivalent in Earth surface environ­ dilution obviously results in reduced abundances, but it does not change
ments, and the observed variability for Eu abundances in the corresponding elemental ratios. The results obtained are plotted using
above-mentioned sample types cannot be explained by a change in its both a linear and a logarithmic scale. The patterns all display different
valence. In practice, all the sedimentary references mentioned above are shapes in the linear scale, but are perfectly parallel in the logarithmic
equally valid, and give similar REE patterns. While the choice of one or scale. This example illustrates the fact that the shapes of the patterns can
any other set of reference values may vary in different laboratories only be compared using a logarithmic scale. The use of a linear scale
worldwide, the most widely used sediment reference for REE is PAAS, should hence be avoided. The only case a linear scale could be used
which corresponds to the average values of a dozen of post-Archean would be when one wants to compare samples with very similar
shale samples from Australia, as initially proposed by Nance and Tay­ normalized abundances. But even there, the usefulness of using a linear
lor (1976) and Taylor and McLennan (1985), and recently scale for the y-axis would be objectively debatable.
re-determined using high-quality REE measurements by Pourmand et al.
(2011). Concentrations of an element X normalized to one of these ref­ 5. Calculation of REE anomalies: La, Ce, Eu, Y
erences are generally noted as Xsn (« sn » standing for « shale normalized
»). Xn is used for normalization with a chondritic reference. One of the interests of REE patterns is to highlight the different
This type of normalization using sediment reference values is also behavior of certain elements relative to their immediate neighbors. This
suited for biological samples, since the REEs they contain are de facto of decoupling can be estimated using the ratio X/X*, where X is the
crustal origin, even for those samples formed in seawater. Instead, the measured concentration of the element, and X* its theoretical concen­
use of any organic material for normalization purpose would most likely tration as estimated assuming a smooth REE pattern (i.e. a pattern in
complicate the interpretation of resulting REE patterns. which three neighboring REEs would be perfectly aligned). A positive
One typical oversight that is made today by some geochemists and anomaly occurs when any element X is in excess in the sample (X/
cosmochemists is the omission of Pm from the other normalized REE X*>1), and a negative anomaly when it is in deficit (X/X*<1).
elements to be plotted on the x-axis. The x-axis does not correspond to a X* can be determined in many different ways, and has been the
list of elements, but to the atomic number (Z) of the elements, which subject of much development in the literature. The reader is referred to
here varies from 57 (La) to 71 (Lu). The omission of Pm (Z = 61) pro­ the work of Lawrence et al. (2006) for a thorough synthesis of possible
duces a jump from Nd (Z = 60) to Sm (Z = 62), and a local change in the equations for estimating X*. There are two philosophies for estimating
shape of the pattern that can be significant. The cause of this mistake is X*: 1/by interpolating or extrapolating X* linearly, which implies
simple. When REE data were initially produced using techniques that reasoning with REE patterns having concentrations represented using a
did not allow all the REEs present in the samples to be determined, linear scale; or 2/by interpolating or extrapolating X* geometrically,
analysts plotted the elements using the correct ‘distances’ between hence considering an y-axis with a logarithmic scale. Both ways are still
them. Nowadays, the use of ICP-MS techniques generally allows all the used by the Earth and environmental sciences community. This issue has
REEs to be analyzed in samples. Since Pm is a radioactive element with a been discussed in detail recently, showing that linear interpolations or
very short half-life (a few years), it is virtually non-existent in nature, extrapolations could lead to erroneous results for REE anomalies (Barrat

6
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

Fig. 6. REY pattern of a coralline algae showing how La*, Ce*, and Eu* are
estimated. Note that the y-axis is logarithmic.

/ /
Ce Ce∗ = Cen x Ndn Pr2 (equation 3)
n

If Pr is not determined, this anomaly can also be estimated by

interpolating Ce* with Nd and Sm. The calculation is made using the
correct distance between Sm and Nd, and is only rigorous if Pm has not
been omitted. The equation is then:
/ /
Ce Ce∗ = Cen x Smn Nd2n (equation 4)

The way to estimate La* is similar, and this concentration can be

Fig. 5. REE patterns of calculated mixtures between a theoretical seawater
sample and pure water plotted using a linear (a) and a logarithmic scale (b). See
extrapolated from Pr and Nd concentrations:
text for discussion. / /
La La∗ = Lan x Nd2n Pr3 (equation 5)
n
et al., 2023). Even if they are still being used routinely in certain fields
(e.g., marine geochemistry), these linear estimations are not rigorous, In the case of REY patterns, the Y anomaly is calculated with the Y/
and should be avoided. Instead, geometric interpolations or extrapola­ Ho ratio because these elements have the same ionic radii (remember
tions should be preferentially used when calculating REE anomalies. that Z-value are on the x-axis, and that Z-value variations are equivalent
Fig. 6 illustrates how La, Ce and Eu anomalies are calculated. The rec­ to ionic radius variation for REEs) This ratio is calculated directly with
ommended equations are listed below using concentrations normalized normalized or non-normalized concentrations. The usual practice is to
to a chondritic reference (Xn), but the same equations also apply for calculate the Y/Ho ratio using raw concentrations expressed in the same
normalization to a shale material (Xsn). unit (i.e., ng/g or μmol/l). As this number is dimensionless, it is
The Eu anomaly is the most straightforward to calculate. Europium is important to specify the units in which the concentrations are expressed.
placed between Sm and Gd. If there is no anomaly for these elements (i.
e., no pollution), Eu* can be estimated using their geometric mean: 6. REE anomalies in polluted samples
/ /
Eu Eu ∗ = Eun (Smn X Gdn )1/2 (equation 1) Two cases of REE pollution can be considered here, depending on
whether the pollutant display a regular REE pattern or not. One example
Similarly, if there is no anomaly in La, we can calculate the anomaly is the case of REE pollution linked to the production of phosphate fer­
in Ce in the same way by interpolating Ce* using the geometric mean of tilizers (e.g., Gaudry et al., 2007). The effluents discharged by this in­
the normalized abundances of La and Pr: dustry produce a non-uniform increase in the concentrations of all the
/ /( )1/2 REEs, without creating anomalies that are unknown in natural samples.
Ce Ce ∗ = Cen Lan xPr (equation 2) In this case, data processing and estimation of the anomalies do not
require any specific calculations. The previous equations apply. How­
n

However, this equation is not recommended due to the frequent ever, human activities may also generate REE pollution related to the
presence of a La anomaly, particularly in marine or oceanic environ­ use of products specifically containing certain purified REEs such as e.g.
ments. Geometric extrapolation of Ce* from Pr and Nd abundances is permanent magnets (Nd, Dy, Pr, Tb), petroleum fluid catalytic cracking
recommended: (FCC) catalysts (La, Ce), polishing (La, Ce), alloys (La, Ce, Pr, Nd, Gd),
rechargeable batteries (light REEs), glasses and ceramics (La, Ce, Pr, Er,

7
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

Y), medical imaging (Gd). The consumption of REEs is growing expo­ concentration in the sample) can be made using the abundances of
nentially and anthropogenic REE pollution is now well documented for neighboring REEs, except for Eu as its relative abundance compared to
Gd (e.g., Bau and Dulski, 1996b), La and Sm (e.g., Kulaksiz and Bau, other rare REEs can fluctuate (see above discussion). We can propose at
2011, 2013). The REE patterns have made it possible to identify them least three ways of estimating Gd* by interpolating it with Sm and Tb
thanks to the resulting anomalies produced. Pollution by other REEs will (equation (6)), by extrapolating from Nd and Sm (equation (7)), or Tb
inevitably be detected in the coming years. To be quantified, the use of and Dy (equation (8)).
REE patterns and the calculation of anomalies (X/X*) will have to be / / ( )
well mastered, particularly in the field of environmental sciences and Gd Gd ∗ = Gdn Sm1/3 x Tb2/3
n (equation 6)
ecotoxicology. Examples of patterns of samples polluted with Gd, La and / /
Sm will be briefly discussed here to illustrate the principles of anomaly Gd Gd∗ = Gdn x Nd2n Smn (equation 7)
calculation, which can be transposed for other REEs in the future. / /
Fig. 7a features a pattern of shell from the Bay of Brest (Patella vul­ Gd Gd∗ = Gdn x Dyn Tb2n (equation 8)
gata; Le Goff et al., 2019), representing an excess Gd anomaly generated Note again that for equation (7), the existence of Pm cannot be
by the release of urine into coastal seawater, following injection of omitted. Depending on the shape of the pattern, equation (8) seems one
Gd-based contrast agents (Le Goff et al., 2019). Gadolinium is practically of the best choices here.
the sole REE contained in these products, although other REEs and Y can Nevertheless, REE pollution may be much more complex. For
be also present in trace amounts (Veiga et al., 2020; Ben Salem and example, the shell of a bivalve taken from the Rhine near Leverkusen
Barrat, 2021). This pollution does not generate other anomalies. Esti­ (Germany; Merschel and Bau, 2015) displays distinctive anomalies in
mation of Gd* concentrations (i.e. the non-anthropogenic Gd both La and Sm (Fig. 7b). Despite its apparent simplicity, this case study
illustrates the difficulties that environmental geochemists will have to
overcome in the next few years due to enhanced REE pollution issues.
First, it is certain that the La excess is here anthropogenic, since this
anomaly is not present in the unpolluted waters of the Rhine (Kulaksiz
and Bau, 2011, 2013). It can be estimated in the same way as above,
using Pr and Nd (equation (5)). The Sm anomaly is also of anthropogenic
origin and probably linked to effluents discharged upstream by the same
industry (Kulaksiz and Bau, 2013). It is possible that this industry also
discharges other REE. However, these additional REE pollutants do not
produce any marked anomalies here, but we cannot rule out the possi­
bility that the anthropogenic contribution is “hidden” but not negligible
elsewhere in the REE pattern, in agreement with Merschel and Bau
(2015). In other words, Sm* must be estimated from elements unaf­
fected by pollution. We could choose elements heavier than Sm to
extrapolate Gd*, but this seems rather uncertain here. We have already
expressed our reservations about Eu. There is also a Gd anomaly here,
and the Rhine is also contaminated with GBCA. We must therefore rely
on the lighter elements, and once again on Pr and Nd, to interpolate
Sm*, but we cannot know a priori whether these elements are not also
affected by pollution. Using these two elements, the Sm anomaly could
be estimated with the following equation:
/ /
Sm Sm∗ = Smn xPr2 Nd3n (equation 9)
n

7. REE patterns as a tool to evaluate data quality

We have summarized above the general principles that have led

geochemists and cosmochemists to adopt REE patterns over the last
decades. The REE distribution patterns in living organisms are also
controlled by the same properties, reflecting the characteristics of their
source material (e.g., seawater, nutrients), hence also justifying their use
by biochemists and ecotoxicologists to present and interpret REE data.
Importantly, these diagrams can be first used to assess the overall quality
of measured data, and to identify results that are not reliable. The Oddo-
Harkins effect is universal and also applies to the distribution of REEs in
organisms, which means that smooth patterns are expected when
normalizing a set of REE values to a chondritic or shale reference, except
for potential REE anomalies in the event of anthropogenic pollution (Gd,
Sm, etc.) or specific microbiological use (e.g., La, Ce, Eu). Living or­
ganisms (and the samples derived from them) are often poor in REEs,
and accurate determination of their abundances can be extremely
Fig. 7. PAAS-normalized REE patterns of (a) a limpet shell (Le Goff et al.,
2019) displaying a huge Gd anomaly (a), and (b) of a Corbicula shell displaying difficult, if not a real analytical challenge. Below, we discuss selected
large La and Sm anomalies (Merschel and Bau, 2015). The red dotted lines show patterns from recently published studies in order to illustrate how
possible ways to estimate La*, Sm* and Gd* values. Note that the y-axis is analytical problems can be detected (Figs. 8 and 9).
logarithmic. (For interpretation of the references to colour in this figure legend, - There is very limited data published so far for soft tissue standards
the reader is referred to the Web version of this article.) and an effort should be made in the near future to develop and

8
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

Fig. 8. a: PAAS-normalized REE patterns of some reference materials prepared Fig. 9. PAAS-normalized REE patterns of selected mushrooms from Poland and
with soft tissues (Danezis et al., 2017); b: PAAS-normalized REE patterns of Yunnan. Data from Mędyk and Falandysz (2022) and Mędyk et al. (2023).
selected ready-to-eat fruit purees (Henríquez-Hernández et al., 2023).
- REEs are also frequently measured in food. Fig. 8b presents selected
characterize reference tissues in order to validate REE data obtained on REE patterns for ready-to-eat baby fruit purees (Henríquez-Hernández
organic materials. Danezis et al. (2017) analyzed two standards, a et al., 2023). The reported REE concentrations are very low, and their
mussel soft tissue and a beef muscle (Fig. 8a). Their results for the mussel determination is therefore particularly challenging. The sample exhib­
tissue (BCR 668) are very satisfactory, and similar to certified values. iting the greatest REE levels in that study exhibits a rather flat
The shape of the obtained pattern is similar to that of bivalve tissues PAAS-normalized pattern. As these purees are essentially made from
measured elsewhere (e.g., Wang et al., 2020; Castro et al., 2023), vegetal products (fruit, sugar) grown on soils derived from chemical
exhibiting for instance a small positive anomaly in La characteristic of a weathering of continental rocks, a shale-like pattern was indeed ex­
seawater origin. On the other hand, the resulting REE pattern identifies pected. However, this pattern shows discrepancies for the heaviest REEs,
small negative anomalies in Ho and Yb that have never been identified most likely of analytical origin. Furthermore, the most REE-depleted
in the marine environment so far. Considering the analytical difficulty of samples show irregular patterns exhibiting hectic anomalies, which
measuring accurate and precise REE abundances for such material, these clearly show that the technique used in that particular study did not
anomalies are probably not real and therefore should not be over­ allow the determination of reliable REE abundances for such low con­
emphasized. Determining REE abundances in the beef muscle standard centration levels. While the results obtained can be used to infer overall
(RM8414) is even more difficult considering that REE concentrations are REE abundances in the order of <10− 4 x PAAS, they do not allow dis­
an order of magnitude lower. The REEs accumulated in a bovine’s cussing the shape of REE pattern in these particular samples. From a
muscles most likely derive from nutritional inputs and additional feed toxicological point of view, the conclusions drawn by the authors are
supplements. Although no indication is available on this subject, it is valid. However, these data should not be included in future databases to
reasonable to assume that the animal was fed mainly with plant-derived discuss pollution or pollution emergence, since they are not sufficiently
products, whose compositional elements ultimately derive from soils. accurate, and since the resulting anomalies are artifacts.
One can presume that the REE patterns for these products must be -The distribution of REEs in mushrooms mostly reflects the compo­
similar to the upper continental crust, hence being most likely similar to sition of soils on which they grow (Zocher et al., 2018). Mędik et al.
a shale reference (e.g., PAAS). Instead, the obtained results show a very (2023) have analyzed numerous chanterelle samples from China and
large positive La anomaly, together with notable enrichments in middle Poland, whose REE patterns show pronounced anomalies, particularly
REEs (Sm, Eu, Gd) and negative anomalies in Dy and Yb. Such REE for heavy REEs (Fig. 9a). Negative Tb and Yb anomalies were obtained
anomalies are unlikely, as they would imply unknown processes of in these samples, but they are most likely erroneous since no negative Tb
biological REE fractionation. The problem highlighted here is probably anomalies have ever been reported in continental crust rocks and
analytical, suggesting that the procedure used in that particular study extraterrestrial material. Ytterbium anomalies are also unknown in
did not allow for REEs in this type of matrix to be quantified with suf­ terrestrial rocks, and have only been encountered in the mineral phases
ficient accuracy at such low concentrations. of a few primitive meteorites formed at very low fO2 (e.g., Crozaz and

9
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

Lundberg, 1995; Barrat et al., 2014). Other analyses of Polish mush­ Data availability
rooms have been carried out by Mędyk and Falandysz (2022) using a
chemical protocol adapted for low-level samples (i.e. after separation of Data will be made available on request.
REEs by ion-exchange chromatography), and the resulting patterns were
correct (Fig. 9b). Acknowledgements
These few examples demonstrate the utility of REE patterns not only
for representing data, but also for detecting analytical problems or We thank Milena Horvat for the editorial handling, the two anony­
limitations. Detecting positive or negative anomalies for REEs in low- mous reviewers for their constructive reviews, and Richard C. Green­
level organic samples requires high-quality measurements, validated wood (Open University) for discussion.
by analyses of well-characterized reference materials. Efforts in this area
are therefore needed, as standards of organic matter are currently References
insufficiently characterized.
Barrat, J.A., Keller, F., Amossé, J., Taylor, R.N., Nesbitt, R.W., Hirata, T., 1996.
Determination of rare earth elements in sixteen silicate reference samples by ICP-MS
8. Perspectives after Tm addition and ion exchange separation. Geostand. Newsl. 20 (1), 133–140.
Barrat, J.A., Zanda, B., Moynier, F., Bollinger, C., Liorzou, C., Bayon, G., 2012.
Geochemistry of CI chondrites: major and trace elements, and Cu and Zn isotopes.
The REE patterns have been the tool of reference for representing
Geochem. Cosmochim. Acta 83, 79–92.
REE data in earth sciences and cosmochemistry for over fifty years. Barrat, J.A., Zanda, B., Jambon, A., Bollinger, C., 2014. The lithophile trace elements in
While the number of studies using REE in ecotoxicology or biology has enstatite chondrites. Geochem. Cosmochim. Acta 128, 71–94.
Barrat, J.A., Dauphas, N., Gillet, P., Bollinger, C., Etoubleau, J., Bischoff, A.,
increased considerably in recent years, researchers in these fields have
Yamaguchi, A., 2016. Evidence from Tm anomalies for non-CI refractory lithophile
not yet fully adopted them. Other representations are used, but have element proportions in terrestrial planets and achondrites. Geochem. Cosmochim.
many drawbacks. While they enable concentration levels to be visual­ Acta 176, 1–17.
ized, they do not allow for a careful assessment of data quality, the Barrat, J.A., Bayon, G., Carney, R.S., Chauvaud, L., 2022a. Rare earth elements as new
biogeochemical proxies in deep-sea mussels. Chem. Geol. https://doi.org/10.1016/j.
determination of anomalies or of any general decoupling of certain el­ chemgeo.2022.121102.
ements relative to other REEs in studied samples. The aim of this paper Barrat, J.A., Chauvaud, L., Olivier, F., Poitevin, P., Bayon, G., Ben Salem, D., 2022b. Rare
was to explain the general principles and main properties of these dia­ earth elements and yttrium in suspension-feeding bivalves (dog cockle, Glycymeris
glycymeris L.): accumulation, vital effects and pollution. Geochem. Cosmochim. Acta
grams, originally developed for rocks, so that they can be better un­ 339, 12–21.
derstood by researchers initiating work on biological materials. Barrat, J.A., Bayon, G., Lalonde, S., 2023. Calculation of cerium and lanthanum
Until recently, it was extremely difficult to obtain reliable data on anomalies in geological and environmental samples. Chem. Geol. 615, 121202
https://doi.org/10.1016/j.chemgeo.2022.121202.
biological materials, given their low concentrations and the cumber­ Bau, M., 1996. Controls on the fractionation of isovalent trace elements in magmatic and
some techniques required to carry out such analyses. The development aqueous systems: evidence from Y/Ho, Zr/Hf, and lanthanide tetrad effect. Contrib.
of ICP-MS has made it possible to obtain high-quality data on this type of Mineral. Petrol. 123, 323–333.
Bau, M., Dulski, P., 1996a. Distribution of yttrium and rare earth elements in the Penge
material, opening up new fields of research. Moreover, with the
and Kuruman iron-formations,Transvaal Supergroup, South Africa. Precambrian Res.
increasing consumption of REEs by the industry and other human ac­ 79, 37–55.
tivities, these elements have now become emerging pollutants whose Bau, M., Dulski, P., 1996b. Anthropogenic origin of positive gadolinium anomalies in
river waters. Earth Planet Sci. Lett. 143, 245–255.
impact on ecosystems and the natural environment needs to be carefully
Bau, M., Möller, P., Dulski, P., 1997. Yttrium and lanthanides in eastern Mediterranean
investigated. seawater and their fractionation during redox-cycling. Mar. Chem. 56, 123–131.
At present, a substantial effort is required to better characterize REE Bau, M., Balan, S., Schmidt, K., Koschinsky, A., 2010. Rare earth elements in mussel
abundances in living organisms and to identify corresponding biological shells of the Mytilidae family as tracers for hidden and fossil high-temperature
hydrothermal systems. Earth Planet Sci. Lett. 299, 310–316.
pathways. In other words, are REEs involved in metabolic reactions or Bau, M., Schmidt, K., Pack, A., Bendel, V., Kraemer, D., 2018. The European Shale: an
not? Do they accumulate in certain organs? The recent discovery of the improved data set for normalisation of rare earth element and yttrium
importance of certain REEs in the metabolism of methanotrophic species concentrations in environmental and biological samples from Europe. Appl.
Geochem. 90, 142–149.
and siderophores has shown that biological fractionation of REE is Bayon, G., Toucanne, S., Skonieczny, C., André, L., Bermell, S., Cheron, S., Dennielou, B.,
possible (e.g., Semrau et al., 2018; Wang et al., 2020; Kraemer and Bau, Etoubleau, J., Freslon, N., Gauchery, T., Germain, Y., Jorry, S., Ménot, G., Monin, L.,
2022), which should be further investigated in future studies. Similarly, Ponzevera, E., Rouget, M.L., Tachikawa, K., Barrat, J.A., 2015. Rare earth elements
and neodymium isotopes in World river sediments revisited. Geochem. Cosmochim.
the mechanism of transfer of the REEs from one trophic level to another, Acta 170, 17–38.
possibly associated with potential accumulation effects and decoupling, Bayon, G., Lemaitre, N., Barrat, J.A., Wang, X., Feng, D., Duperron, S., 2020. Microbial
must also be evaluated (e.g., Marginson et al., 2023), together with the utilization of rare earth elements at cold seeps related to aerobic methane oxidation.
Chem. Geol. 555, 119832.
possible impact of biological and microbiological processes on the
Bellefroid, E.J., Hood, A.V.S., Hoffman, P.F., Thomas, M.D., Reinhard, C.T.,
oceanic REE cycle or on their distribution in soils and terrestrial surface Planavsky, N.J., 2018. Constraints on Paleoproterozoic atmospheric oxygen levels.
environments. These studies are also directly relevant to various envi­ Proc. Natl. Acad. Sci. USA 115, 8104–8109.
Ben Salem, D., Barrat, J.A., 2021. Determination of rare earth elements in gadolinium-
ronmental issues regarding the impact of pollution on ecosystems, hence
based contrast agents by ICP-MS. Talanta 221, 121589.
the need to further develop ecotoxicological studies in order to identify Brewer, A., Dror, I., Berkowitz, B., 2022. Electronic waste as a source of rare earth
emerging REE pollutants. For these reasons, REE patterns will certainly element pollution: leaching, transport in porous media, and the effects of
become an essential tool for researchers working on these issues. nanoparticles. Chemosphere 287, 132217.
Castro, L., Farkas, J., Munro Jenssen, B., Piarulli, S., Ciesielski, T.M., 2023.
Biomonitoring of rare earth elements in Southern Norway: distribution,
CRediT authorship contribution statement fractionation, and accumulation patterns in the marine bivalves Mytilus spp. and
Tapes spp. Environ. Pollut. 335, 122300.
Charles, C., Barrat, J.A., Pelleter, E., 2021. Trace element determinations in Fe-Mn oxides
Jean-Alix Barrat: Writing – original draft, Conceptualization. Ger­ by high resolution ICP-ms after Tm addition. Talanta, 122446.
main Bayon: Writing – original draft, Conceptualization. Chayla, B., Jaffrezic, H., Joron, J.L., 1973. Analyse par activation dans les neutrons
épithermiques: application à la détermination d’éléments traces dans les roches.
C. R. Acad. Sci. Paris 277, 273–275.
Coryell, C.D., Chase, J.W., Winchester, J.W., 1963. A procedure for geochemical
Declaration of competing interest interpretation of terrestrial rare-earth abundance patterns. J. Geophys. Res. 68,
559–566.
The authors declare that they have no known competing financial Crozaz, G., Lundberg, L.L., 1995. The origin of oldhamite in unequilibrated enstatite
chondrites. Geochem. Cosmochim. Acta 59, 3817–3831.
interests or personal relationships that could have appeared to influence
the work reported in this paper.

10
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

Danezis, G.P., Pappas, A.C., Zoidis, E., Papadomichelakis, G., Hadjigeorgiou, I., standardization (ID-IS) and multi-stage internal standardisation. Geostand. Geoanal.
Zhang, P., Brusic, V., Georgiou, C.A., 2017. Game meat authentication through rare Res. 30, 245–271.
earth elements fingerprinting. Anal. Chim. Acta 991, 46–57, 2017. Marginson, H., MacMillan, G.A., Grant, E., Gérin-Lajoie, J., Amyot, M., 2023. Rare earth
Danezis, G.P., Zoidis, E., Zhang, P., Pappas, A.C., Tsagkaris, A.S., Papachristidis, C.A., element bioaccumulation and cerium anomalies in biota from the Eastern Canadian
Papadomichelakis, G., Hadjigeorgiou, I., Georgiou, C.A., 2019. Tissue distribution of subarctic (Nunavik). Sci. Total Environ. 879, 163024.
rare earth elements in wild, commercial and backyard rabbits. Meat Sci. 153, 45–50. Massari, S., Ruberti, M., 2013. Rare earth elements as critical raw materials: focus on
Dang, D.H., Ma, L., Ha, Q.K., Wang, W., 2022. A multi-tracer approach to disentangle international markets and future strategies. Resour. Pol. 38, 36–43.
anthropogenic emissions from natural processes in the St. Lawrence River and Masuda, A., 1962. Regularities in variation of relative abundances of Lanthanide
Estuary. Water Res. 219, 118588. elements and an attempt to analyse separation index patterns of some minerals.
De Baar, H.J.W., Bacon, M.P., Brewer, P.G., Bruland, K.W., 1985. Rare earth elements in J. Earth Sci. Nagoya Univ. 10, 173–187.
the pacific and atlantic oceans. Geochem. Cosmochim. Acta 49, 1943–1959. Mędyk, M., Falandysz, J., 2022. Occurrence, bio-concentration and distribution of rare
Elderfield, H., Upstill-Goddard, R., Sholkovitz, E.R., 1990. The rare earth elements in earth elements in wild mushrooms. Sci. Total Environ. 851, 158159.
rivers, estuaries, and coastal seas and their significance to the composition of ocean Mędyk, M., Falandysz, J., Chidi Nnorom, I., 2023. Scandium, yttrium, and lanthanide
waters. Geochem. Cosmochim. Acta 54, 971–991. occurrence in Cantharellus cibarius and C. minor mushrooms. Environ. Sci. Pollut.
Garcia-Solsona, E., Jeandel, C., Labatut, M., Lacan, F., Vance, D., Chavagnac, V., Res. 30, 41473–41484.
Pradoux, C., 2014. Rare earth elements and Nd isotopes tracing water mass mixing Merschel, G., Bau, M., 2015. Rare earth elements in the aragonitic shell of freshwater
and particle-seawater interactions in the SE Atlantic. Geochem. Cosmochim. Acta mussel Corbicula fluminea and the bioavailability of anthropogenic lanthanum,
125, 351–372. samarium and gadolinium in river water. Sci. Total Environ. 533, 91–101.
Gast, P.W., Hubbard, N.J., Wiesmann, H., 1970. Chemical composition and petrogenesis Minster, J.F., Allègre, C.J., 1978. Systematic use of trace elements in igneous processes.
of basalts from Tranquillity base. Proc. First Lunar Sci. Conf., Geochim. Cosmochim. Contrib. Mineral. Petrol. 68, 37–52.
Acta, Supp. 2 (1), 1143–1163. Nance, W.B., Taylor, S.R., 1976. Rare earth element patterns and crustal evolution–I.
Gaudry, A., Zeroual, S., Gaie-Levrel, F., Moskura, M., Boujrhal, F.Z., El Moursli, R.C., Australian post-Archean sedimentary rocks. Geochem. Cosmochim. Acta 40,
Guessous, A., Mouradi, A., Givernaud, T., Delmas, R., 2007. Heavy metals pollution 1539–1551.
of the Atlantic marine environment by the Moroccan phosphate industry, as Noack, C.W., Dzombak, D.A., Karamalidis, A.K., 2014. Rare earth element distributions
observed through their bioaccumulation in Ulva lactuca. Water Air Soil Pollut. 178, and trends in natural waters with a focus on groundwater. Environ. Sci. Technol. 48,
267–285. 4317–4326.
German, C.R., Elderfield, H., 1990. Application of the Ce anomaly as a paleoredox Nozaki, Y., Zhang, J., Amakawa, H., 1997. The fractionation between Y and Ho in the
indicator: the ground rules. Paleoceanography 5, 823–833. marine environment. Earth Planet Sci. Lett. 148, 329–340.
Goldberg, E.D., 1961. In: Oceanography, M. Sears (Ed.), Chemistry in the Oceans, vol. Ouyang, J., Wang, X., Zhao, B., Yuan, X., Yuchun Wang, Y., 2003. Effects of rare earth
67. Am. Assoc. Adv. Sci. Publ., pp. 583–597 elements on the growth of Cistanche deserticola cells and the production of
Goldschmidt, V.M., 1958. Geochemistry. Oxford Univ. Press, London, p. 730. phenylethanoid glycosides. J. Biotechnol. 102, 129–134.
Grenier, M., Garcia-Solsona, E., Lemaitre, N., Trull, T.W., Bouvier, V., Nonnotte, P., Ozaki, T., Enomoto, S., Minai, Y., Ambe, S., Ambe, F., Tominaga, T., 1997. Determination
Souhaut, M., Lacan, F., Jeandel, C., 2018. Differentiating lithogenic supplies, water of lanthanides and other trace elements in ferns by instrumental neutron activation
mass transport, and biological processes on and off the Kerguelen Plateau using rare analysis. J. Radioanal. Nucl. Chem. 217, 117–124.
earth element concentrations and neodymium isotopic compositions. Front. Mar. Sci. Pack, A., Russell, S.S., Shelley, J.M.G., van Zuilen, M., 2007. Geo- and cosmochemistry of
5, 426. the twin elements yttrium and holmium. Geochem. Cosmochim. Acta 71,
Gromet, L.P., Haskin, L.A., Korotev, R.L., Dymek, R.F., 1984. The ‘north American shale 4592–4608.
composite’: its compilation, major and trace element characteristics. Geochem. Pol, A., Barends, T.R.M., Dietl, A., Khadem, A.F., Eygensteyn, J., Jetten, M.S.M., Op den
Cosmochim. Acta 48, 2469–2482. Camp, H.J.M., 2014. Rare earth metals are essential for methanotrophic life in
Hatch, G.P., 2012. Dynamics in the global market for rare earths. Elements 8, 341–346. volcanic mudpots. Environ. Microbiol. 16, 255–264.
Henderson, P., 1984. General geochemical properties and abundances of the rare earth Pourmand, A., Dauphas, N., Ireland, T.J., 2011. A novel extraction chromatography and
elements. In: Henderson, P. (Ed.), Rare Earth Element Geochemistry. Elsevier, MC-ICP-MS technique for rapid analysis of REE, Sc and Y: revising CI-chondrite and
pp. l–32. Post-Archean Australian Shale (PAAS) abundances. Chem. Geol. 291, 38–54.
Henríquez-Hernández, L.A., Acosta-Dacal, A.C., Boada, L.D., Zumbado, M., Serra- Pourret, O., van der Ent, A., Hursthouse, A., Irawan, D.E., Liu, H., Wiche, O., 2022. The
Majem, L., Luzardo, O.P., 2023. Concentration of essential, toxic, and rare earth ‘europium anomaly’ in plants: facts and fiction. Plant Soil 476, 721–728.
elements in ready-to-eat baby purees from the Spanish market. Nutrients 15, 3251. Prudêncio, M.I., Valente, T., Marques, R., Sequeira Braga, M.A., Pamplona, J., 2015.
Jahn, B., Pol, A., Lumpe, H., Barends, T.R., Dietl, A., Hogendoorn, C., Op den Camp, H.J. Geochemistry of rare earth elements in a passive treatment system built for acid
M., Daumann, L.J., 2018. Similar but not the same: first kinetic and structural mine drainage remediation. Chemosphere 138, 691–700.
analyses of a methanol dehydrogenase containing a europium ion in the active site. Schnetzler, C.C., Philpotts, J.A., 1968. In: Ahrens, L.H. (Ed.), Partition Coefficients of
Chembiochem 19, 1147–1153. Rare-Earth Elements and Barium between Igneous Matrix Material and Rock-
Jarvis, I., Jarvis, K.E., 1992. Plasma spectrometry in the earth sciences: techniques, Forming-Mineral Phenocrysts, -I in Origin and Distribution of the Elements, vol. 30.
applications and future trends. Chem. Geol. 95, 1–33. Pergamon Press, pp. 929–937.
Jenner, G.A., Longerich, H.P., Jackson, S.E., Fryer, B.J., 1990. ICP-MS -A powerful tool Semrau, J.D., DiSpirito, A.A., Gu, W., Yoon, S., 2018. Metals and methanotrophy. Appl.
for high-precision trace element analysis in Earth sciences: evidence from analysis of Environ. Microbiol. 84.
selected U.S.G.S. reference samples. Chem. Geol. 83, 133–148. Shannon, R.D., 1976. Revised effective ionic radii and systematic studies of interatomic
Johannesson, K.H., Stetzenbach, K.J., Hodge, V.F., 1997. Rare earth elements as distances in halides and chalcogenides. Acta Crystallogr., Sect. A 32, 751–767.
geochemical tracers of regional groundwater mixing. Geochem. Cosmochim. Acta Shiller, A.M., Chan, E.W., Joung, D.J., Redmond, M.C., Kessler, J.D., 2017. Light rare
61, 3605–3618. earth element depletion during Deepwater Horizon blowout methanotrophy. Sci.
Jochum, K.P., Weis, U., Schwager, B., Stoll, B., Wilson, S.A., Haug, G.H., Andreae, M.O., Rep. 7, 10389.
Enzweiler, J., 2016. Reference values following ISO guidelines for frequently Taylor, S., McLennan, S., 1985. The Continental Crust: its Composition and Evolution.
requested rock reference materials. Geostand. Geoanal. Res. 40, 333–350. Blackwell Scientific Publications, p. 312.
Kamber, B.S., Webb, G.E., 2001. The geochemistry of late Archaean microbial carbonate: Tostevin, R., Shields, G.A., Tarbuck, G.M., He, T., Clarkson, M.O., Wood, R.A., 2016.
implications for ocean chemistry and continental erosion history. Geochem. Effective use of cerium anomalies as a redox proxy in carbonate-dominated marine
Cosmochim. Acta 65, 2509–2525. settings. Chem. Geol. 438, 146–162.
Kamber, B.S., Greig, A., Collerson, K.D., 2005. A new estimate for the composition of Veiga, M., Mattiazzi, P., de Gois, J.S., Nascimento, P.C., Borges, D.L.G., Bohrer, D., 2020.
weathered young upper continental crust from alluvial sediments, Queensland, Presence of other rare earth metals in gadolinium-based contrast agents. Talanta
Australia. Geochem. Cosmochim. Acta 69, 1041–1058. 216, 120940.
Karamalidis, A.K., Torres, S.G., Hakala, J.A., Shao, H., Cantrell, K.J., Carrol, S., 2013. Wallace, M.W., Hood, A.S., Shuster, A., Greig, A., Planavsky, N.J., Reed, C.P., 2017.
Trace metal source terms in carbon sequestration environments. Environ. Sci. Oxygenation history of the Neoproterozoic to early Phanerozoic and the rise of land
Technol. 47, 322–329. plants. Earth Planet Sci. Lett. 466, 12–19.
Kraemer, D., Bau, M., 2022. Siderophores and the formation of cerium anomalies in Walsh, J.N., Buckley, F., Barker, J., 1981. The simultaneous determination of the rare
anoxic environments. Geochem. Persp. Let 22, 50–55. earth elements in rocks using inductively coupled plasma source spectrometry.
Kulaksiz, S., Bau, M., 2011. Rare earth elements in the Rhine River, Germany: first case Chem. Geol. 33, 141–153.
of anthropogenic lanthanum as a dissolved microcontaminant in the hydrosphere. Wang, X., Barrat, J.A., Bayon, G., Chauvaud, L., Feng, D., 2020. Lanthanum anomalies as
Environ. Int. 37, 973–979. fingerprints of methanotrophy. Geochem. Persp. Let. 14, 26–30.
Kulaksiz, S., Bau, M., 2013. Anthropogenic dissolved and colloid/nanoparticle-bound Watkins, P.J., Nolan, J., 1992. Determination of rare earth elements, yttrium, scandium
samarium, lanthanum and gadolinium in the Rhine River and the impending and hafnium using cation-exchange separation and inductively coupled plasma-
destruction of the natural rare earth element distribution in rivers. Earth Planet Sci. atomic emission spectrometry. Chem. Geol. 95, 131–139.
Lett. 362, 43–50. Wilkin, R.T., Lee, T.R., Ludwig, R.D., Wadler, C., Brandon, W., Mueller, B., Davis, E.,
Lawrence, M.G., Greig, A., Collerson, K.D., Kamber, B.S., 2006. Rare earth element and Luce, D., Edwards, T., 2021. REEs as natural tracers for in situ remediation of
yttrium variability in south east queensland waterways. Aquat. Geochem. 12, 39–72. groundwater. Environ. Sci. Technol. 55, 1251–1259.
Le Goff, S., Barrat, J.A., Chauvaud, L., Paulet, Y.M., Gueguen, B., Ben Salem, D., 2019.
Compound-specific recording of gadolinium pollution in coastal waters by great
scallops. Sci. Rep. 9, 8015.
Makishima, A., Nakamura, E., 2006. Determination of major, minor and trace elements
in silicate samples by ICP-QMS and ICP-SF-MS applying isotope dilution-internal

11
J.-A. Barrat and G. Bayon Chemosphere 352 (2024) 141487

Zepeda, V.K., Kamber, B.S., Ghidan, O.Y.A., 2023. Direct accurate Eu anomaly analysis in Zocher, A.L., Klimpel, F., Kraemer, D., Bau, M., 2021. Assessing the bioavailability of
very high Ba/Eu silicate samples by triple-quadrupole ICP-MS in MS/MS mass shift dissolved rare earths and other trace elements: digestion experiments with aquatic
mode. Chem. Geol. https://doi.org/10.1016/j.chemgeo.2023.121827. plant species Lemna minor (“duckweed” reference standard BCR-670). Appl.
Zocher, A.L., Kraemer, D., Merschel, G., Bau, M., 2018. Distribution of major and trace Geochem. 134, 105025 https://doi.org/10.1016/j.apgeochem.2021.105025.
elements in the bolete mushroom Suillus luteus and the bioavailability of rare earth
elements. Chem. Geol. 483, 491–500.

12