Mesures quantitatives par ICP-MS:

Applications en mode liquide et par ablation laser

Olivier Bruguier – Géochimiste

Géosciences Montpellier

Séminaire Pôle Balard, 31 Mars 2016

Séminaire Pôle Balard

A quoi sert… l’ICP-MS?

Responsables Scientifiques D. BOSCH & J.L. SEIDEL

Responsables Techniques: O. BRUGUIER & R. FREYDIER

S. DELPOUX & C. DOUCHET

France

International

Pyrénées-Méditerranée

X7 series Element XR

Agilent 7700x

ICAP-Q

Laser Excimer

ICP-MS: Inductively Coupled Plasma Mass Spectrometry

Technique récente (1980)

è analyses multi-élémentaires, rapides, faibles LoD

0.00 0.05 0.10 0.15 0.200

10000

20000

30000

40000

50000

cps

Concentration (ppb)

Tm

U

BEC = 0.079 ppt

BEC = 0.022 ppt

r^2 = 0.9999

r^2 = 0.

9997

Forte sensibilité (> 2 x 106 cps/ppb)

Gamme dynamique élevée (de qques cps à 109 cps)

The sampling sites S1 (on the roof of the geological institute ata height of about 30 m in the center of Strasbourg) and BKG wereestablished as “permanent stations”. K1 and A1 (in private gardens)were established as non-permanent stations. S1, K1 and A1 aresituated west, south and northeast, respectively of the industrialnorth harbor (Fig. 1). The non-permanent sampling site K2 islocated southeast of the industrial north harbor.

2.2. Sampling method

The applied passive sampling technique (Grobéty et al., 2010)uses a Sigma-2 sampler to collect the ambient-air particles withsizes in the range of 2.5e80 mm, which are mainly deposited viasedimentation into a small acceptor dish (w5.5 cm in diameter). A

four-week sampling period guarantees accumulation of sufficientairborne material in rural environments for both trace element andPb, Sr and Nd isotope measurements, whereas one or two weeks ofsampling yields enough dust material in industrial areas or close tobusy highways. In this study, a three to mostly four-week collectionperiod has been adopted for all stations in order to have similarsampling conditions for all sites. One exception is the traffic site inFreiburg (TR), for which only a one-week collection period wasadopted for each sample (SI Table S1).

2.3. Acid digestion

Each of the dust samples was transferred from the acceptor dishinto Teflon flask vessels with 10e20 mL of HNO3 (1 N) and then

Fig. 1. Map showing sampling sites of Sigma-2 collectors (red circles) and industrial parks (triangles). Also given are the principal wind directions, as indicated by the wind rose.CWI: chemical waste incinerator, DWI: domestic waste incinerator, TPP: thermal power plant, SP: steel plant, BHPS: biomass heating power station, OR: oil refinery, PP: paperproducer. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

F. Guéguen et al. / Atmospheric Environment 62 (2012) 631e645 633

dried on a hot plate. Subsequently, the sample was digested in 1 mLof HNO3 sub-boiled and 1mL of HF Suprapur at 70 !C during 2 days.After drying, 1 mL of HClO4 was added to the sample, which wasthen heated to 150 !C, where it was kept for one day, and evapo-rated. 20% of the sample was taken to determine major and traceelement concentrations using inductively-coupled plasma atomicemission spectroscopy (ICP-AES) and inductively-coupled plasmamass spectroscopy (ICP-MS). Uncertainties of elemental concen-tration are <5% (Steinmann and Stille, 1997). The weight of theairborne particles collected on the small acceptor dish within theSigma-2 passive sampler device was too small for precise weighing.Therefore, the trace element concentrations are not given in ppm(mg g"1) but as elemental mass deposited in the Sigma-2 dish ona specific surface per day (ng m"2 d"1). The results are given in thesupporting information (SI Table S4).

2.4. Analysis of Sr, Nd, Pb isotope ratios

Strontium, Pb and Nd were separated from other elements byextraction chromatography on Eichrom Sr Spec, TRU Spec and LnSpec resins (Pin and Santos Zalduegui, 1997; Deniel and Pin, 2001).The chemical separation was validated by analyzing commonlyused standards, including NBS 981, NBS 987, an in-house Nd stan-dard, and BCR-1. The isotopic compositions of Sr, Pb and Nd weremeasured with a Neptune-Thermo Scientific multi-collector (MC)ICP-MS under dry plasma conditions using a membrane desolvator(Apex, CPI International). The NBS 987 standard measured duringthe same period yields a 87Sr/86Sr ratio of 0.71027 # 0.00006 (2s,n $ 19). 143Nd/144Nd isotopic ratio determined on an in-house Ndstandard yields 0.51172 # 0.00002 (2s, n $ 29) corresponding toa value of 0.51184 % 0.00001 for the La Jolla standard. The143Nd/144Nd isotopic ratios are expressed as 3Nd values (DePaoloand Wasserburg, 1976). For Pb isotopic ratios, instrumental massfractionation was monitored and corrected online using a SRM997 Tl standard (White et al., 2000). The NBS 981 Pb standardyielded mean values of 208Pb/206Pb $ 2.1661 # 0.0004 (2s, n $ 12),206Pb/204Pb $ 16.931 # 0.004 (2s, n $ 12) and206Pb/207Pb $ 1.0935 # 0.0001 (2s, n $ 12). The procedural blanks

were 0.35 ng, 0.125 ng and 0.005 ng for Pb, Sr and Nd, respectively.The results are given in SI Table S4.

3. Results and discussion

3.1. Site-dependent trace metal depositions

The mean mass deposition rates (MDRs) have been calculatedfor the bulk PM collected at four stations during 8e21 months (A1:n $ 8; S1: n $ 20; K1: n $ 9; BKG: n $ 19) as well as at differentstations to the North of the cities in the industrial harborcollected during shorter periods (H: H1eH8; S3: n $ 14) andclose to high-traffic highways (TR: n $ 4; S2: n $ 3) (SI Table S1).Fig. 2A shows themean “total”MDRs (mgm"2 y"1), as derived fromthe major (Mg, Al, Ca and Fe) and trace element (V, Cr, Cu, Mn, Co,Ni, Cu, Zn, As, Rb, Sr, Y, Zr, Mo, Cd, Sn, Sb, Cs, Ba, La, REE, Pb, U andTh) data for the PM samples collected at these sites. Of course, theseMDR values are minimum values because not all of the elementshave been analyzed and because the data are presented inelemental form rather than in their more probable oxide, sulfate, orsilicate forms. Nevertheless, the data clearly reveal that theMDRs ofthe PM from traffic and industrial emissions are at a rather similarlevel, which is substantially higher than those at the sites K1, S1,and A1, and w10 times higher than at the BKG station. The MDRvalues at sampling site S1, located on the roof of the institute in thecenter of Strasbourg, appear to be slightly lower than at samplingsite K1, situated at 1.5 m above ground level and closer to theindustrial harbor and traffic. It is of note that sampling site A1(village of Auenheim), situated in a private garden at a distance of<3 km NE of the steel plant (SP), has considerably lower MDRs thansite K1, situated 3.5 km SW of the SP (Fig. 1).

Fig. 2B shows the relative variations of MDRs for individualelements in comparison with the most enriched average samplesfrom the harbor (100%) and the traffic (100%). Important major andtrace elements from industrial emissions in the harbor (H) are Pb,Fe, Cr, Co, Mo, Cd, Ni, As, Zn and Sb, with highest mean MDRs (allvalues given in: mg m"2 y"1) observed for the metals Fe (671), Cr(109), Ni (73.5), Zn (70.4), and Pb (2.93). Themost important traffic-

A B

Fig. 2. A: Mass deposition rates for traffic (TR), harbor (H), city centers of Strasbourg (S1) and Kehl (K1), the village of Auenheim (A1), and the Vosges mountains (BKG), calculatedfrom the following major element and trace element data: Mg, Al, Ca and Fe, and V, Cr, Cu, Mn, Co, Ni, Cu, Zn, As, Rb, Sr, Y, Zr, Mo, Cd, Sn, Sb, Cs, Ba, La, REE, Pb, U and Th (SI Table S4).B: spider diagram showing trace metal concentrations in dust samples collected from the same sites as in (A). Each of the most enriched elements of the samples has always beenset equal to 100%.

F. Guéguen et al. / Atmospheric Environment 62 (2012) 631e645634

Caractérisation chimique des particules atmosphériques en milieu urbains, industriels et ruraux

(Gueguen et al., 2012)

Fig. 3. Ternary diagrams for LaCeSm (A), LaCeV (B), CdLaMo (C), MoCoPb (D), ThCrU (E), and CaFeAl (F). The greyish fields are defined from tree-bark biomonitoring (Lahd Geageaet al., 2007, 2008a). Fertilizer values are from this study (annotated (I) and (II) (SI Table S3)), and from Hayumbu et al. (1995) and Otero et al. (2005). The white fields defined by a redline in (A) and (B) correspond to PM2.5 and PM10 compositions fromMexico City (Moreno et al., 2008). (For interpretation of the references to colour in this figure legend, the readeris referred to the web version of this article.)

F. Guéguen et al. / Atmospheric Environment 62 (2012) 631e645636

Fig. 5. Excess lead fluxes through time: Gazon-du-Faing (a) and Rossely (b) in the Vosges Moutains; Kohlhütte Moor (c) in the Black Forest; Etang de la Gruère (d) in the Swiss JuraMountains; Port des Lamberts (e) in the Morvan massif. The isotopic 206Pb/207Pb ratios of excess lead are reported on the graphs. Attention must be paid to the Rossely diagramwhich presents a double Y-axis scale.

17B. Forel et al. / Journal of Geochemical Exploration 107 (2010) 9–20

Analyses des éléments en traces dans les tourbières

(Forel et al., 2010)

Identification des principales sources de

pollution

changes are therefore able to supply indirect evidence, provided thatthey have been well preserved over time. Ombrotrophic peatlandsmay fulfil such requirements. They are almost exclusively fed by theatmosphere so that mineral matter inputs are expected to be low(Shotyk, 1996a,b) and they are supposed to be hydrologically isolatedfrom the substratum. Finally, anthropogenic contributions can bemore easily extracted from the raw geochemical signal, and henceinterpreted in historical terms (Shotyk et al., 1998, 2003;Mighall et al.,2002; Kylander et al., 2005; Jouffroy-Bapicot et al., 2007; see alsoBindler, 2006 for a review). Among all metals studied, lead is often the

most informative because its affinity with organic matter tends toprevent post-depositionalmigration (Dumontet et al., 1990; Vile et al.,1999; Shotyk et al., 1997, 2003; Weiss et al., 1999). Lead was notnecessarily exploited by prehistoric societies, but because it isassociated to many ore deposits including copper, it can be regardedas a proxy for various types of mining.

In the present study, we aim to establish a continuous history ofmetal-related activities in the Vosges massif, from the first signs ofmetallurgy to the most recent periods. The exceptional richness ofarchaeological and historical records in the region is a considerable

Fig. 1. (a) Geographical situation of the two sites studied, together with other peatlands previously studied (Shotyk et al., 1998; Monna et al., 2004a; Le Roux et al., 2005); (b) Map ofVosges Mountains and sampling sites. The 206Pb/207Pb ratios of some Pb ore deposits measured by Marcoux (1987) are also reported.

10 B. Forel et al. / Journal of Geochemical Exploration 107 (2010) 9–20

Laser Compex 102 -  Laser à gaz (ArF), pulsé (15ns) -  Energie: 200 mJ -  Fréquence de tir: 1 à 20 Hz -  Diamètre d’analyse: 5µm à 160µm -  Vitesse de déplacement: 3 à 20 µm/s

Element XR -  Monocollecteur -  Secteur Magnétique -  Li à U : 150 ms Mixing Cell

Plasma

ConesIon Optics

Quadrupole

Detector

Computer

Ablation Cell

Figure Modified From D. Sinclair, 1999

Laser

Flux d’Hélium

Plasma

Optique onique

Filtre de masse (Quad or Sector field)

Ordinateur

Cellule d’Ablation Cellule de mélange

Détecteur

•  Forte résolution spatiale •  Ablation de la plupart des matériaux •  Forte sensibilité

C sys tème pe rme t tan t une amplification cohérente de la lumière par émission stimulée;

C Premier laser à rubis (T. Maiman, 1960).

Couplage Laser / ICP-MS:

Effets thermiques (fusion, vaporisation, production d’atomes et d’ions, plasma) ET

Mecaniques (microfracturation, éjection de fragments)

Energie des photons ➠ matériel cible

Sylvester & Kosler, 2003

Processus destructif (avantage/inconvénient)?

50 µm

Applications Carbonates: spéléothèmes, foraminifères, otolithes… Silicates: minéraux majeurs et accessoires… Phosphates: coprolithes, pellets… Verres, Céramiques, Alliages, Métaux…

c. 30µm depth

100 µm

Comparaison Sonde Ionique / Ablation Laser

Problem High predation on mussels

Gilthead Sea bream?

Large individuals (>1Kg) caught in automn-winter in Thau lagoon:

What is the Gilthead Sea bream migration behaviour?

ICP-MS Applications

Mouillot, D., Darnaude, A. Ferraton, F. Mercier L., MARBEC

ICP-MS Applications

Thau lagoon

Mediterranean Sea

-0,050

0,000

0,050

0,100

0,150

0,200

0,250

0,300

0,350

0,400

0,00 500,00 1000,00 1500,00 2000,00

Distance from the nucleus (µm)

Pb (m

g/g)

Spring Automn

Spring Automn

[Pb] ppm Individus d’élevage

Supposed behaviour

Spawning in winter (Sea)

Juvenils come into the lagoon in spring Go out the lagoon in automn

Along the coast in winter

Growth in summer

ICP-MS Applications

Selective zircon accumulation in a new benthic foraminifer,Psammophaga zirconia, sp. nov.A. SABBATINI , 1 A . NEGRI , 1 A . BARTOLINI , 2 C . MORIGI , 3 O. BOUDOUMA,4 E. DINELLI , 5

F . FLORINDO,6 R . GALEAZZI , 1 M. HOLZMANN,7 P . C. LURCOCK,6 L . MASSACCESI ,1

J . PAWLOWSKI7 AND S. ROCCHI3

1Dipartimento di Scienze della Vita e dell’Ambiente Di.S.V.A., Universit!a Politecnica delle Marche, Ancona, Italy2Centre de Recherche sur la Pal"eobiodiversit"e et les Pal"eoenvironnements, UMR 7207 CNRS MNHN UPMC, Mus"eumNational d’Histoire Naturelle, Paris Cedex 05, France3Dipartimento di Scienze della Terra, Universit!a di Pisa, Pisa, Italy4ISTEP (UMR 7193), Universit"e Pierre et Marie Curie, Paris Cedex 05, France5Dipartimento di Scienze Biologiche, Geologiche e Ambientali, Universit!a di Bologna, Bologna, Italy6Istituto Nazionale di Geofisica e Vulcanologia (INGV), Roma, Italy7Department of Genetics and Evolution, Universit"e de Gen!eve, Geneva 4, Switzerland

ABSTRACT

Benthic foraminifera are single-celled eukaryotes that make a protective organic, agglutinated or calcareous

test. Some agglutinated, single-chambered taxa, including Psammophaga Arnold, 1982, retain mineral par-

ticles in their cytoplasm, but the selective mechanism of accumulation is not clear. Here, we report the abil-

ity of a foraminiferal species to select and accumulate zircons and other heavy minerals in their cytoplasm.

In particular, the use of Scanning Electron Microscope coupled with an Energy Dispersive X-ray microanaly-

sis system (SEM–EDS) enabled a representative overview of the mineral diversity and showed that the anal-

ysed Psammophaga zirconia sp. nov. individuals contained dominantly crystals of zircon (51%), titanium

oxides (27%), and ilmenite (11%) along with minor magnetite and other minerals. The studied specimens

occur in the shallow central Adriatic Sea where the sediment has a content of zircon below 1% and of

other heavy minerals below 4%. For that reason we hypothesize that: (i) P. zirconia may be able to chemi-

cally select minerals, specifically zircon and rutile; (ii) the chemical mechanism allowing the selection is

based on electrostatic interaction, and it could work also for agglutinated foraminifera (whether for inges-

tion, like Xenophyophores, or incorporation in the test as in many other described taxa). In particular, this

aptitude for high preferential uptake and differential ingestion or retention of zircon is reported here for

the first time, together with the selection of other heavy minerals already described in members of the

genus Psammophaga. They are generally counted among early foraminifera, constructing a morphologically

simple test with a single chamber. Our molecular phylogenetic study confirms that P. zirconia is a new

species, genetically distinctive from other Psammophaga, and occurs in the Adriatic as well as in the Black Sea.

Received 20 May 2015; accepted 18 January 2016

Corresponding author: A. Sabbatini. Tel.: +39 0712204329; fax: +39 0712204650; e-mail: a.sabbatini@

univpm.it

INTRODUCTION

The ability of monothalamous (single-chambered) foramini-

fera, including Psammophaga, to selectively incorporate min-

eral grains into their cytoplasm has been known for some

decades (Arnold, 1982; Pawlowski & Majewski, 2011; Bal-

lero et al., 2013). The etymology of the name Psammophaga

reflects the ingestion of sediment grains, which are retained

in the cytoplasm as inclusions. Traditionally this genus

belongs to soft-walled monothalamous foraminifera, and

Pawlowski et al. (2003a) proposed a Precambrian origin for

this group. In the molecular phylogeny of foraminifera

(Pawlowski et al., 2002a, 2013) the genus Psammophaga is

included in clade E of the monothalamids. The extraordinary

© 2016 John Wiley & Sons Ltd 1

Geobiology (2016) DOI: 10.1111/gbi.12179

molecular diversity of monothalamous foraminifera has been

documented by Majewski et al. (2007), Habura (2008) and

Gooday et al. (2011). Several Psammophaga species have

recently been described based on molecular and morpholog-

ical characters (Pawlowski & Majewski, 2011; Ballero et al.,2013). One of these species, Psammophaga magnetica,found in shallow-water Antarctic sediments, selectively

ingests magnetite, titanoferous magnetite, and other detrital

minerals such as feldspar and quartz grains, concentrating

them close to the aperture (Pawlowski & Majewski, 2011).

Another species, Psammophaga sapela, collected from salt

marshes and mudflats along the coast of Georgia, U.S.A.,

ingests and stores different mineral phases, including ilme-

nite and zircon (Ballero et al., 2013). Both papers suggest

that these two species are able to select and accumulate

heavy minerals as ballast, although neither study quantified

these minerals in the cytoplasm of Psammophaga and com-

pared them to the mineral composition of the sediments

where they were found.

Faunal studies of living benthic foraminifera from the

central Adriatic Sea have revealed the seasonal occurrence

of individuals of Psammophaga sp. (Sabbatini et al., 2012).Fluctuations in their density apparently reflect an oppor-

tunistic response to pulses of high-quality organic carbon,

suggesting a possible role for the species as a shallow-water

benthic eutrophication indicator. However, as stated

above, the types of minerals ingested by these foraminifera

in the Adriatic Sea have not been investigated previously.

The mineral composition of central Adriatic Sea sedi-

ments reflects the composition of the drainage basins of

the adjacent rivers. The Po River is the most important,

and its drainage basin includes the western and central Alps

and the Ligurian-Emilian portion of the northern Apen-

nines. As a consequence, the main mineral phases found in

the central Adriatic sediments are calcite, dolomite, quartz,

clay minerals (muscovite-illite, chlorites, and smectite), and

feldspars, with occasional occurrences of amphiboles and

serpentine (Dinelli & Lucchini, 1999). Heavy minerals like

ilmenite, rutile, and zircon are generally present, but in

very low abundance (Pigorini, 1968; Brondi et al., 1979).The above cited literature, reporting on mineral inges-

tion by individuals in different environmental settings,

stimulated us to investigate the contents of Psammophagafrom the Adriatic Sea in order to explore the mineralogical

phase selection and possible cellular mechanisms underly-

ing this phenomenon.

MATERIALS AND METHODS

Analyses of the ingested minerals in Psammophaga

Microscopic analyses, mineralogy, and petrography

Replicated samples (n = 5) for this study were collected

from five transects, each with three stations, perpendicular

to the coast in the central Adriatic Sea in 12–17 m water

depth using a Van Veen grab (Figs 1 and 2 and S1). The

specimens were isolated from the upper 2–3 cm of surface

sediments of three replicates of each station. Immediately

after recovery, the sediment samples were gently washed,

and living Psammophaga individuals were picked from the

>90 lm residues; 23 selected specimens were prepared for

Scanning Electron Microscopy (SEM).

The remaining sediment residues were fixed in 10% for-

malin buffered with sodium borate, with 2 mL of Rose

Bengal solution (1 g L!1) for population studies of living

stained foraminiferal assemblages. A total of 291 Psam-mophaga specimens were picked.

Living specimens were photographed after isolation

using a Nikon Eclipse E 600 POL stereomicroscope with

transmitted light. Scanning Electron Microscope micro-

graphs were obtained using a ZEISS SUPRA 55 VP SEM

with a 3rd generation GEMINI field emission column,

allowing a spatial image resolution down to 1.0 nm, in

scattered (SE) and backscattered (BSE) modes, at the

University of Paris VI-UMPC.

Chemical compositions of mineral grains inside 17 Psam-mophaga specimens were analysed using an energy disper-

sive X-ray (EDS) microanalysis system (SAHARA Silicon

Drift Detector with PGT’s Spirit Software, Rocky Hill, NY,

USA) allowing high counting rates, and related cartogra-

phies were produced. Hyperspectral X-ray images were pro-

duced using the SAHARA Silicon Drift Detector with the

same analytical conditions (15 kV accelerating voltage,

7 mm working distance, and a beam current of 8 nA).

Adriatic Sea

12

3A

B

45

6 CD

E78

9

1011

12

131415

Italy

Ancona

Oce

an d

ata

view

Fig. 1 Location map.

© 2016 John Wiley & Sons Ltd

2 A. SABBATINI et al.

The size, shape, and chemical composition of the crystals

found inside 4 Psammophaga individuals were determined

using a Philips XL30 SEM with SE and BSE detectors for

imaging and EDS for analysis at Dipartimento di Scienze

della Terra, Universit!a di Pisa. The individuals were broken

and smeared onto a glass slide and carbon coated. A total

115 crystals inside Psammophaga were imaged by SEM–SEand BSE detectors, collecting size/shape data for every

crystal analysed by SEM–EDS. Furthermore, a volume

evaluation of the heavy mineral grains inside four

Psammophaga cytoplasms was performed using the

measured dimensions of the crystals (see Table S1).

Magnetic analyses

The magnetic properties of Psammophaga individuals were

investigated using a Princeton Measurements Corporation

MicroMag 2900 alternating gradient magnetometer

(AGM) at the Istituto Nazionale di Geofisica e Vulcanolo-

gia (INGV) in Rome. We chose to use an AGM (rather

than a vibrating sample magnetometer) to maximize sensi-

tivity and avoid potential inaccuracies caused by movement

of Psammophaga individuals during sample vibration.

We measured remanent magnetization during stepwise

acquisition of an isothermal remanent magnetization

(IRM) up to 1 tesla (T); the remanent coercive force

(Bcr), evaluated by stepwise back-field application on the

saturation IRM; hysteresis loops; and first-order reversal

curves (FORCs). We performed FORC analyses (Roberts

et al., 2014) to characterize magnetostatic interactions

and magnetic domain state distributions in the studied

samples. Analyses were made with a field increment of

2 mT, Hu ranging from !60 to +60 mT, Hc from 0 to

120 mT, an averaging time of 100 ms, and N = 137

measured FORCs. Data were processed, smoothed, and

plotted using the FORCINEL program (Harrison & Fein-

berg, 2008). A smoothing factor of 5 was applied to data.

Finally, we evaluated the possible contribution of super-

paramagnetic grains (SP grains) to the total remanence

using the method of Wang et al. (2010), which works by

monitoring the viscous decay of an applied IRM. Assum-

ing that the room-temperature viscous decay of the IRM

100 s after its application is due to thermal relaxation of

the magnetization carried by the SP fraction, the SP parti-

cle percentage can be estimated as MRS(SP)% = 100

(MRS0 ! MRS100)/MRS0, where MRS0 is the remanence

measured immediately after application of a 1 T saturating

field and MRS100 is the remanence measured 100 s after

field application.

Sediment analyses: sedimentology, petrography, andgeochemistry

The collected samples were used for grain-size analysis of

surface sediments. The sample locations are given in Fig. 1.

Grain-size analyses were carried out on wet sediment sam-

ples pre-treated with H2O2-16 vol. solution to remove

organic matter. The coarser fractions (>63 lm) of samples

were analysed by sieving using the sifting machine

FRITSCH analysette 3 (Spartan pulverisette 0) while the

finer fractions (<63 lm) were analysed by X-ray sedigraph

(Micrometrics 5100). Before sieving, the samples were

dried to determine their dry bulk masses, and 25 g of the

dry sediment was removed for bulk sediment measure-

ments. Sieving was performed with a vibration amplitude

of 1.5 mm and the sieving time was 20 min in total. Eight

grain-size fractions from >500 to >63 lm (See Table S2)

100 µm 100 µm

100 µm100 µm

A B

C D

Fig. 2 Type specimens for Psammophaga

zirconia sp. nov. preserved in 10% formalin

from the Central Adriatic Sea. (A–D) Holotype

and paratype from station D11 (Portonovo,

Central Adriatic Sea). In (A) holotype

specimen (DISVA UPM 2015-1) of 375 lmlength; in (B) paratype specimen (DISVA UPM

2015-2) of 275 lm length. In (C) and (D)

details of the terminal aperture, single and

flexible at the end of a very short neck. The

wall, thin and transparent with a shiny

surface, is also visible. (A) Light microscope

photograph of holotype DISVA UPM 2015-1,

375 lm length. (B) Light microscope

photograph of paratype DISVA UPM 2015-2,

275 lm length. (C) Light microscope

photograph of holotype DISVA UPM 2015-1.

(D) Light microscope photograph of paratype

DISVA UPM 2015-2.

© 2016 John Wiley & Sons Ltd

Heavy minerals uptake in foraminifera 3

Marche, Ancona (Italy) under the following registration

numbers: holotype DISVA UPM 2015-1 (Fig. 2A,C);

paratypes: DISVA UPM 2015-2 (Fig. 2B,D) and DISVA

UPM 2015-3a-c (Fig. S1A,D,G).

A total of 291 specimens, fixed in 10% formalin buffered

with sodium borate, with 2 mL of Rose Bengal solution

(1 g L!1) were analysed for the morphological and popula-

tion studies.

Diagnosis

Test free, single-chambered, pyriform, elongate, or spheri-

cal in shape with a single simple terminal aperture; the wall

thin and transparent, appears to be mainly organic with a

sparse surface dusting of fine particles. The cell body does

not entirely fill test lumen; abundant mineral inclusion

retained throughout the entire cytoplasm.

Description

Pyriform, elongate or spherical theca with length usually

ranging ~0.20–0.70 mm. The apertural (proximal) end is

usually rounded and tapering at the distal end toward a

large, simple aperture (Figs 2 and 3C). The wall is thin

and transparent with a shiny surface. The aperture is clearly

visible, single and flexible which may occur at the end of a

very short neck; it seems not to be prolonged into any

internal structure (Fig. 2C). The aperture is presumed to

be sufficiently plastic to allow for the passage of sediment

grains of various sizes. Where visible the cytoplasm is white

and fine grained and often drawn out into a point immedi-

ately inside the aperture. The cytoplasm contains numerous

cytoplasmic inclusions of mineral grains that typically

appear to fill the entire test (Figs 2 and 3C). Pseudopodia

reticulate solely from a short peduncle that emerges from

the aperture through which food and sediment grains also

pass. While sorting both living and Rose Bengal stained

specimens, strong brightness of Psammophaga specimens

was noted together with magnetism of some analysed indi-

viduals. As is typical of the genus, P. zirconia ingests and

retains abundant mineral inclusions and appears to prefer

heavier minerals, in particular zircon and rutile.

Molecular features

Eight partial SSU rDNA sequences were obtained from 3

P. zirconia specimens collected at the type locality (Fig. 4).

The sequences were submitted to GenBank (Accession

numbers LN886765 to LN886772). Sequence length

ranges from 877 to 1024 nt, the GC content ranges from

46.2% to 46.8%. Phylogenetic analysis of the eight

sequences groups them into a strongly supported clade

(96% BV).

Distribution

Living P. zirconia was found at 15 sites from the central

Adriatic Sea, between 12 and 17 mwd (Fig. 1). It was

commonly found in low numbers, <0.5 specimens per

10 cm!2, but at one location (D11) at 14.5 mwd it

occurred in great abundance (>150 specimens per

10 cm!2) following seasonal input of organic matter. Fau-

nal studies of living benthic foraminifera from the central

Adriatic Sea have described the occurrence, seasonally

100 µm 100 µm

500 µm 100 µm

A B

C D

Fig. 3 Stereomicroscope photographs of living

Psammophaga zirconia individuals. (A) Track

of the P. zirconia movement in the sediment;

arrow indicates the final position of the

specimen. (B) The appearance of P. zirconia in

the sediment where it lives. (C) Living

P. zirconia; intracellular mineral inclusions are

visible. (D) Pseudopodial activity of P. zirconia.

© 2016 John Wiley & Sons Ltd

Heavy minerals uptake in foraminifera 5

Foraminifères benthiques (test carbonaté ou agglutiné)

Accumulation de particules minérales dans le cytoplasme

predominantly mineral inclusions concentrated around the

apertural end and they seem to be more widely spread in

P. sapela and P. crystallifera. In all observed individuals of

the new species, cytoplasmic mineral grains appear to fill

the test entirely (Figs 2, 3 and 5 and S2). Although P. zir-conia shares some morphological peculiarities with all

Psammophaga species, it differs in mineral inclusion fea-

tures. The new species incorporates heavy minerals and

predominantly selects equidimensional zircon, while

P. magnetica ingests magnetite along with minor amounts

of other lighter minerals; in other Psammophaga species,

authors described a mixture of mineral grains, most of

them heavy. As for P. zirconia, the presence of zircon,

ilmenite and Ti oxides was also reported in P. sapela,although their quantitative proportion in the cytoplasm

remains undescribed. If left isolated in a culture dish with

no movement or extraneous sediment overnight, P. zirco-nia does not egest most of its mineral inclusion as

observed in P. sapela (Ballero et al., 2013). P. zirconiainhabits geographically and ecologically distinct settings

with respect to P. magnetica, dominant in polar habitats,

and P. sapela, in marsh and mudflat areas.

Phylogenetically, P. zirconia branches at the base of a

clade containing P. crystallifera and Psammophaga sp. from

Denmark, but the branching is not supported (Fig. 4).

The six sequences obtained from the Adriatic samples are

very similar to two unpublished sequences from the Black

Sea, collected near Sevastopol (Gooday et al., 2011).

Together all these sequences form a strongly supported

(96% BV) clade, which is considered here as corresponding

to the new species. It has to be highlighted that the genetic

diversity of the genus Psammophaga is very high. Our phy-

logenetic analyses group all Psammophaga sequences into

17 clades, each of which probably corresponds to a separate

species. However, none of these clades, including all

described species, is closely related to P. zirconia, reinforc-ing the arguments for describing it as a new species.

RESULTS

The SEM and EDS analyses performed on the cytoplasm

of P. zirconia specimens allowed us to identify the pres-

ence of different minerals and also the shape and composi-

tion (and hence the nature) of the mineral crystals. These

analyses revealed that all analysed specimens are filled with

heavy minerals among which zircon particles are largely

dominant. The total volume of crystals throughout the

Psammophaga cytoplasm reaches more than 70%, as also

supported by SEM photographs and related EDS cartogra-

phies (Figs 5 and S2). Quantitative examination performed

on four previously broken Psammophaga individuals by

SEM and EDS showed that the mineral grains comprised

51% nearly equidimensional zircons, 27% Ti oxides (mainly

rutile), 11% ilmenite, and 3% magnetite, along with minor

amounts of other lighter minerals (Fig. 6). Aggregates

(10 9 20 lm) of lm-sized halite cubes have been

observed, which are possibly linked to precipitation from

individuals stained with Rose Bengal solution. Once the

numbers of crystals are converted to volumes, more than

80% is represented by zircon, with only a mere 20% by

other heavy mineral grains (Ti oxides, ilmenite, and mag-

netite; Fig. 5; see Table S1 and Fig. S2). The magnetic

analyses, performed on cells of P. zirconia, consistently

indicated the presence of magnetite (Fig. S3). More than

90% of saturation magnetization was reached in a field of

200 mT and the coercivity of remanence is in the range of

A B

C D

Fig. 5 SEM photographs and related EDS

cartographies of living Psammophaga zirconia

specimen. (A) SEM BSE photograph of

P. zirconia filled with mineral grains. (B)

Combined color image of three elements

(Red = Zr, Green = Ti, Blue = Si) from an

elemental analysis of the same specimen in

(A). Minerals in pink are zircons, in green Ti

oxides, in blue silicates and quartz. (C) Results

of differential elemental analysis mapping

(Red = Fe, Green = Ti, Blue = Si). Identified

minerals in red are Fe oxides, in orange

ilmenite, in green rutile, in blue silicates and

quartz. (D) Zircon crystals inside the cell of

P. zirconia; different specimen from (A–C).

Images resolution: 1024 9 768 9 16 bits.

© 2016 John Wiley & Sons Ltd

Heavy minerals uptake in foraminifera 7

15 mT. The FORC diagram has a peak near the origin

with open contours that diverge toward the Hu axis; this is

consistent with the presence of low-coercivity (fine) mul-

tidomain magnetite grains (MD grains) (2–10 lm) which,

in turn, is consistent with the size of the intracellular Fe

oxides revealed from elemental analysis by EDS (Fig. S2).

However, the absence of a viscous decay of magnetization

(Wang et al., 2010) seems to exclude the concomitant

presence of ultrafine magnetite fractions (cubic SP,

20–30 nm).

On the basis of these results we further explored the

grain size distribution and chemical composition of the

sediments in the sites where individuals were collected, in

order to understand possible relationships between

P. zirconia density and sediment characteristics.

Our results show that the grain-size distribution in sur-

face samples is mostly clayey silt and silt (Shepard, 1954).

Coarser sediments (sand >30%) are dominant near-shore at

shallow depths and at the southernmost stations. Sites

located further away from the coast have a more abundant

silt fraction, while northernmost stations are richer in clay

(Table S2).

Geochemical analyses results indicate that silicate and

quartz are the dominant minerals in all samples ranging

from 31.93 (%) to 42.88 (%) in the bulk sediment and

34.11 (%) to 43.39 (%) in the finer fraction (<63 lm)

(Table S3). Zr concentrations range from 56 to 194 lg g!1

(10–183 ppm) and its value is higher (50–328 lg g!1;

126–329 ppm) in the finer fraction (<63 lm) of the studied

samples with increasing values to the south of the studied

transects (Fig. 1). The results for Ti (1320–3540 lg g!1)

and Fe (15 180–36 440 lg g!1) display a mainly increasing

trend with distance from the coastline (Table S3).

Additionally SEM and EDS investigations performed on

bulk sediment samples, on the separate coarsest (>63 lm)

and finest (<63 lm) fractions, reveal that silicates and

quartz dominate (Fig. S4B,C), while zircon, ilmenite, Fe

and Ti oxides are rare minerals. In detail, in the analysed

finest fraction of the sites A3, D12 and E15, 1% of miner-

als are zircon, 4% ilmenite, 4% Ti oxides (mainly rutile)

and 3% Fe oxides (magnetite) (Fig. 6).

DISCUSSION

Our data point to an unusually high concentration of

zircon minerals in the P. zirconia cytoplasm, which con-

trasts with the extreme rarity of Zr in the sediment,

strongly suggesting a selectivity of mineral ingestion in

these organisms (Figs 6 and S4).

To a lesser extent, but more significantly, Ti–Fe-bearingminerals such as ilmenite, Ti oxides, and magnetite are also

accumulated along with the zircon crystals; they all are rare

minerals in the sediment and constitute the densest mineral

component within a sediment in which silicates and quartz

crystals dominate (Figs 6 and S4). The Zr is associated

only with zircon and no other mineralogical phases, sug-

gesting a very low abundance of this mineral in the sedi-

ment. In the case of Ti, however, the mineralogical phases

are various: Ti is associated with rutile, ilmenite, and titan-

ite and Fe with magnetite, ilmenite, and pyrite. Moreover,

our mineralogical distributions of sediment are consistent

with data published by Spagnoli et al. (2014) for the Cen-

tral Adriatic Sea, reporting low concentrations of Zr, Ti,

and Fe oxides, and a dominance of silicates and quartz.

Significantly, zircon and Ti oxides are both selectively

concentrated in the cell (respectively 51% and 27%, vs. 1%

and 4%), whereas Fe oxides (magnetite) and ilmenite show

cytoplasmic concentrations comparable to those found in

the sediment (Fig. 6). Here we address the question of

how foraminifera can select minerals, in order to under-

stand the possible advantages that this behavior confers in

a morphologically simple organism such as P. zirconia.

19

82

8

0

20

40

60

80

100% Volume of minerals in Psammophaga zirconiaA

3

11

51

27

1 1 16

0

15

30

45

60% Number of crystals in Psammophaga zirconiaB

5.5 3.0

3 41 2 1

4 5 58

1 2

10

3

52

0

15

30

45

60% Number of crystals in the sediment (<63 �m)C

Mineral density (g cm–3)

Fig. 6 (A) Percentage of volume that heavy minerals occupy in the 4

Psammophaga zirconia cells. (B) Percentage of minerals calculated on the

basis of counted crystals inside 4 P. zirconia specimens. (C) Percentage dis-

tribution of the 16 minerals in three sediment samples of the <63 lm frac-

tionin the 900-point analyses. Size and nature of the mineral crystals are

determined by SEM–EDS. Arrow indicates the mineral density increase.

© 2016 John Wiley & Sons Ltd

8 A. SABBATINI et al.

Mécanismes de sélection ? Ingestion ou rétention préférentielle?

Intérêts: ballast? Camouflage? Source de chaleur??

Projet PALMES (CPER): Spectromètres de masse de type MC-ICP-MS et IRMS + couplages – Premiers appareillages de ce type installés en Région Languedoc - Roussillon

Recherche et Diffusion des Connaissances Formation et Apprentissage

Eau Santé Ecologie/Environnement

Transect LA-ICPMS sur otolithe

Sr/Ca = traceur de salinité

→ histoire migratoire des poissons...

0

2

4

6

8

10

12

0 0.5 1 1.5 2 2.5

Distance from otolith nucleus (mm)

Sr88

:Ca4

3 (10

^-3)

Entrée à 1-2 ans + migrant occasionnel

Prestations/Développements pour entreprises

Développements de nouveaux outils d’expertise et de diagnostic