The comparative behaviors of yttrium and ... - Wiley Online Library

23 downloads 39 Views 396KB Size Report
GEOPHYSICAL RESEARCH LETTERS, VOL. 21, NO. 24, PAGES 2677-2680, DECEMBER 1, 1994. The comparative behaviors of Yttrium and. Lanthanides in ...
GEOPHYSICAL RESEARCH LETTERS, VOL. 21, NO. 24, PAGES 2677-2680, DECEMBER 1, 1994

The comparative behaviors of Yttrium and Lanthanides in the seawater of the North Pacific J.Zhang,H. Amakawaand Y. Nozaki Ocean Research Institute,University of Tokyo,Nakano-ku, Tokyo,Japan

Abstract: Yttrium has long been recognized as an eka- interactionsof trace metals. In particular, the comparative lanthanide, becauseof its chemicalcontiguityrelative to the Y/REE observationswill provide useful insights into the chemistries of rare earthelementswhich are, in recentyears, chemicalcharacterof the organicsurfacecoatingsof particles intensivelyutilized for elucidation of metal scavenging which are believed to occur ubiquitously. Few attempts,

processes in theocean.Here,we present thefirstdetailed however,have been made to determineyttrium in seawater. depth-profile of Y(III) in seawatertogetherwith the

Since the historicalreport of Hogdahl et al. /1968] who obtained!30-150 pmol/kg for Y(III) in the deep water of the Y(iII) range70-290 pmol/kgand showthe "nutrient-like" North Atlantic basedon a tediousneutronactivationanalysis, profile bestresembling thatof Ho(III) amongst theotherrare therehas been,to our bestknowledge,no publishedliterature earthelements.The resultsagree well with an expectation on the seawaterconcentrationof yttrium until recently. This based on the similarityin their ionic radii andhencestability is surprisingin contrastto the very precisedeterminationsof constantsof complexation with carbonate ions. Yet the REEs that have been made in various parts of the oceans Ho(m)/Y(ra) ratios in seawatersystematicallyincreasewith [Piepgras and Jacobsen, 1992; Sholkovitz and Schneider, depth,suggesting that Y and Ho are fractionatedduring 1991; Bertramand Elderfield,1993]. Thisis, at leastin part, method for the scavenging by naturalmarineparticulates. This is likely to because there was no conventional resultfrom the differentcomplexationbehaviorin that Y(iiI) determination of yttrium in seawater, and therefore the is moreweakly complexedthan Ho(iI!) with soft organic developmentof new techniqueshouldsupplementthis lack of !igands on thesurfaceof particulatematterduringscavenging knowledgeof Y(III) geochemisrty.Shabaniet aI. [1990] made inthesurface waterbut,oncereleasedintoseawater in thedeep this effort by using inductively coupled plasma mass sea, Y(III) is complexedwith carbonateions equally or spectrometry(ICP-MS), but they determined only a few stronger thanHo. The patternof deepwaterenrichment in the seawatersamples. Clearly, it is not possible to draw any lanthanideseries appears to be consistent with the recent patternof the oceanicY(III) distributionasyet. Therefore,we observation of partitioningbetweensuspended particlesand measuredY(III) in water profile samplescollectedfrom the seawater.Our precisemeasurements alsoindicatedthat Pr and KuroshioCurrentregimeof the westernNorth Pacific. lanthanides in the North Pacific Ocean. The concentrationsof

Tb bestresemble Nd and Dy, respectivelyin their oceanic behavior,whereas Ho and Tm are intermediate between their

neighboring rare earth elements.

Methods

Water sampleswerecollectedat 29ø05'N,142ø5I'E of theIzuOgasawara Trench(waterdepth,9500m)by usingR/V TanseiMaru on September 12, 1993. Samplingwasmadeby usinga

Introduction

Because of thesimplicity andpredictability of theirchemical standardCTD/Niskin-Rossetearray down to 3500m. Unfilterd systematics, rare earth elements (REEs) have received seawaterwas acidifiedto pH95%)asdetermined byusing chemical leachares of suspended particles andseawater and indiumasa yieldmonitor[Shabaniet al., !990]. The isotopes

are89y 139La ' i40Ce' !41pr' observed thesystematic difference in REE partitioningusedforthedeterminatin between seawater andsuspended particulate matter whichis 146Nd, 147Sm, 151Eu, 159Gd, i59Tb, !63Dy, 165Ho,

apparently controlled byadsorption/desorption equilibria. 167Er, 169Tm, 173yb and175Lu. These isotopes arefreeof

Because ofthewell-known chemical similarity between Y(III)- isobaric interferences andnocorrections arerequired. Other andREEs, themeasurements of Y(III) in theoceanicisotopes such as,143Nd, 153Eu, 166Er and172yb also give environment can furtherthe studyof particle/solution results ingood agreement withthose based ontheabove, and therefore theinterferences of BaO+ andREEO + arenegligible.

Copyright 1994bytheAmerican Geophysical Union.

The standardsolutionwasprovidedfrom SPEX Industries,Inc. which is accurate to ñ 0.5 % for each REE.

The Y andREE datawereobtainedby using~1000ml of unfilteredseawater.The fractionof particulate form,however, mustbe negligiblysmalljudgingfrom the caseof REEs [< 4

Paper number 94GL02404 0094-8534/94/94GL.02404503.00 2677

2678

ZHANG ET. AL.: COMPARATIVE

BEHAVIORS

OF YTTRIUM

AND LANTHANIDES

%, Sholkovitz et al., 1994]. The blanks for the entire 2. If theratiois constant with depth,thentherewouldbeno procedure and reagents were measuredin parallel with the fractionationbetweenthe two elementsin their oceanic samplesand suitablycorrected.They were the highestfor the behavior.Suchis thecasefor Tb/Dy so thattheprofileofTo light rare earth elements(LREEs) being, in the surfacewater, is identical to thatof Dy butnotto thatof Gd. Likewise, pt Nd morethanLa. In thecaseof Ho andTin,onthe --.10% for La, Pr and Nd, but relatively minor (< 5 %) for resembles yttriumand the middleand heavyrare earthelements(MREEs other hand, their ratios to the neighboring REEs vary with depth,showingan almost"mirror" & HREEs). We do not reportherethenumbersof Ce dueto the systematically between high blanks (up to -60 % of the signal),presumablydue to image(Figure2). This impliesthatHo is intermediate contaminationfrom the room atmosphere. Five replicate Dy and Er, and Tm between Er and Yb in their oceanic

of determinations of the deepwatercollectedfrom~2000m gave behavior,and hencetheir ratiosmay be usefulas tracers the relative standarddeviationsof the meansrangingfrom + 5 different water masses. The first detailedprofileof yttriumis shownin Figure 1 % for thelightestREE (La) to q-3 % for theheaviestREE (Lu) and Y01I).

together withthoseof theREEs.Theconcentration gradually

Results

increases from70 pmol•g at thesurfaceto 290pmol/kgatthe depthof 3500 m showingso called "nutrient-like" profile.

and

Discussion

TheYOII) is highlycorrelated withall theREEs(presumably

Vertical distributionsof Y and REEs are plottedin Figure 1. Our resultsof La, Nd, Sm, Eu, Gd, Dy, Er, Yb andLu agreewith the profile data of Piepgras and Jacobsen [1992] and Klinkhammer et al. [1983] for the similar locationsthat were obtained by isotope dilution and thermal ionization mass spectrometry (ID-TIMS). Our resultsalsoincludethedataof Pt, Tb, Ho and Tm that are unable to be measuredby ID-TIMS method. For theseelements,the only previousdata axethose of DeBaar et al. [1985] for the easternNorth Pacific basedon neutron activation analysis. Our REE concentrationsare systematically -30% lower than theirs for the same depth hirizons, presumablydue to the differencein the locations,but our depth-profilesare much smootherthan theks. The detailedprofilesof Pr, Tb, Ho andTm obtainedherehave enabled us, for the first time, to make a meaningful comparison with those of their neighboring trivalent REEs. This is done by taking the elementalratios as shownin Figure 0

100

200

3000

20

40

600

2

4

6

8

o; ....

3o..4o

except Ce)withthecorrelation coefficients, R2>0.9.When we plottedthe correlationcoefficientversusthe trivalentREEs

in orderof increasingatomicnumber,a very interesting pattern envisagedas shown in Figure 3a. The correlation

coefficients (R2) of La andtheREEsheavier thanHowith respectto Y(III) are higherthan0.99 and are indistinguishable

fromeach other, whereas theR2 systematically decreases from Ho to Pr with decreasingatomicnumber. Inspectionof theYREE correlation diagrams (althoughnot shown) indicatesthat the failure of goodnessin the Y/REE correlationis notcaused

by random scatter but the concavity of data points particularlyfor the deep water. This suggeststhat thepattern

of R2 is a reflection of thedifference in thegeochemical behavior in that Y(III) more resemblesthe HREEs thanLREEs. The unexpectedsimilarity of the oceanic distributionbetween Y and La may be related to the closestneighboringin the3A

column of periodic table and the absenceof 4f electrons in their electron orbital. By contrast,the very good correlation between Y(III) and the HREEs may be explainedby the similarity in their ionic radii and their strong complexati0n behavior with carbonateions that are the plausibleinorganic solutionligandsfor thosemetals in seawater. 0.1 0

2O2468100 .

I

0.2 0.3 • . i ._--_ ....

0.4

0.12 ,

0.16 0.20 i - _- ,-

0.24

2 1'

b

2 Nd /Gd

3 Pr/La • •/

0O, 4 8 120 I 2 3 4 m

E

..• v

rO.

4 0.24 0'

0.28 ,

..

H

ß

0

4

8

120

1

0.32

0.36

,

.

r

0.16

0.12

0.20

I'

ß

2

Figure 1. The verticalprofilesof Y and rare earthelements Figure2. The atomicratiosof Pr, Tb, Ho andTm relative to in the westernNorth Pacific. All units are givenin pmol/kg. their neighboring trivalentREEsversusdepth. The Note that the profiles deviate from each other particularlyin correlation coefficients (R2) for all thepaired elements are the salinity minimum zone centeredat 750 m.

better than 0.99except forthePr/La pair(R2=0.943).

ZHANGET. AL.: COMPARATIVE BEHAVIORSOF YTTRIUMAND LANTHANIDES Sincethecorrelationcoefficientsof La andHREEs against

2679

(Ho/Y)x100

Y(I]2)areextremely high,it isnotpossible to identify oneof

0.8

the elementsthat has the best resemblanceto yttrium. Therefore, wetakeanotherapproach by definingthedeepwater enrichment factor,EDW = C3000mlCsurface,whereC3000m andCsurfaceare the concentration at 3000m (arbitrarily chosen for thedeepwater)andin thesurfacemixedlayer(< 100

0.9

1.0

1.1

0

1000

m),respectively. Thedeepwaterenrichment factoris plotted

against individual lanthanides withincreasing atomic number inFigure 3b,whichshowsa parabolic curvewithitsminimum

,...,,- 2000

at Tb. The similar pattern can also be seen in the data of

Piepgras andJacobsen [1992].Clearly,themiddleREEsshow theleastdeepwaterenrichment.This suggests thatthe LREEs andHREEsaremorestronglycomplexedwith seawaterligands andhencelessparticle-reactive as comparedto MREEs. By analogy to othertracemetals,one can imaginethat MREEs

3000 S&

somewhat resembleCu whereasLREEs and HREEs behavemore

.

4000

34.0 likeZn andNi duringscavenging by particles[Bruland,1980]. 34.5 35.0 The adsorption/desorption reactions on the suspended Salinity particlesis consideredto play a key role in the REE fracfionation in the ocean,and there appearsto be general Figure 4. The vertical profile of Ho/Y atomic ratio in the

consistencies among these observations. In comparisonof Y(III) with the REEs in their oceanic

profile,theperfectmatch is expectedwhenEDW (Y)= EDW

NorthPacific.The anomalously highratiosin the surface water and the North Pacific intermediate water around the

(REE) andR2=1.TheEDWforY(III)is4.0which issmaller

salinity minimum are hatched. Our error estimatesfor the

Ho/Y ratiosbasedon thedatagivenin Table1 andunpublished thananyof the REEs. This meansthat the oceanicdistribution ones are less than + 4 %.

of Y(III)

is not exactly the same as those of REEs.

Nevertheless, the middle REEs between Sm and Ho have the

EDW valuesof 4.5 + 0.2 whichare closeenoughto showthe distributionsimilar to that of Y(III).

0.99 in the diagramof Figure3a and EDW = 4.5 :t:0.2 areto be satisfied,then only Ho remainsamongstthe other REEs. Althoughthesecriteriaare somewhatarbitrary,we conclude

.•_ 0.96.

,•0.94. I

that Y(III) in seawaterbest resemblesthe distributionof Ho in the ocean. This is not unreasonablebecauseHo has an ionic

'• O.92

0.9o]• , . ,.-..,......

....... , . ,....... , .... ,

La Ce Pr Nd PmSmEu Od Tb Dy Ito Er Tm Yb Lu

radiussimilarto thatof Y(III) and their stabilityconstants of complexation with carbonate are approximately the same[Lee and Byrne, 1993]. The atomicratio of Ho/Y in seawatershowsa decreasefrom

10-



(b)

ß

Thiswork

x

the surfaceto a subsurface minimumof 0.85 x 10-2 at 200m

andthen asystematic increase to1.06x 10-2at3000m (Figure

• xPiepgras & Jacobsen. 1992•

4). Thesurface maximum is associated withthelow salinity water. Somewhat highvaluesabovethegeneraltrendareseen

aroundthe salinityminimumlayer (700-1000m in depth).

These suggest thatthefractionation between Y(III) andHo(and otherREEs)is takingplacenot only in the watercolumnbut •

-,

-,

-,

-,.1

ß 1'.i

- ß - ß ß , . ,.•..t-.,

l.• Ce Pt NdPmSmEuGd 'lb •

. ,

Ho ExTmYb Lu

also in the sourceregions of distinct water masses. In the previousworks[e.g. Klinkhammeret al., 1983; DeBaaret al.,

1985], theoceanic distributions of REEsareoftencompared withthatof dissolved silicaconcentration. Although thereare Figure 3a.Thepattern of correlation coefficients (R2) of generalsimilarities,it is clearfrom theabovearguments that Y(III)versus trivalent REEs withincreasing atomic number.thereis no geochemical significance in suchcomparison. On Element

One mayargue thatthispattern maybea reflection of the

the other hand, Y(III)

serves as an eka-lanthanide which is

increased uncertainty ofmeasurements fromtheHREEs toward guite meaningfullyused for normalizationof REEs in theLREEs, butthegood correlation between LaandY rules out geochemical andoceanographic interpretations. thepossibility.

their

The meanY(III)/Ho(III) ratio in seawateris about100,

3b.The pattern ofthedeep water enrichment factor (EDW) whereas thoseof pelagicsediments [Turekian andWedepohl,

versus trivalent REEs(seetext).Thedataof Piepgras and 1961; Rakinand Glasby,1979] and shale[Haskinand Haskin, Jacobsen (1992) arebased onthestation TPS271-1in the 1966;NanceandTaylor,1976] are40 and47, respectively.

western North Pacific gyre.Ourdatacanbefittedwellwiththe Sincethe pelagicsedimentsand shaleare thoughtto be curve ofEDW= 0.0729(6.45- Z)2 + 4.3where Z istheatomic materials of ultimateremovalfrom the oceans,Y(III) and number of REEs,withR2 -- 0.964. Ho(III) mustbe fractionated eitherduringweathering on the

2680

ZHANG ET. AL.: COMPARATIVE BEHAVIORS OF YTTRIUM AND LANTHANIDES

continentsor duringscavenging in the watercolumnand diagenesis in the underlying sediments. It seems likely that Y(III) in seawateris lessparticle-reactive thanHo andthe fractionation duringparticlescavenging is responsible for the higherY(III)/Ho(III) ratio observed in seawater thanin the abovemarinedeposits.ByrneandLee, [1993]indicates that thecomplexargon behaviorof Y(III) withsoftorganicligands best resembles that of Sm but not Ho, and hence the

Elderfield, H.,andM.J. Greaves, Therareearth elements in seawater, Nature 296, 214-219, 1982.

Goldberg, E. D.,M. Koide,R. A. Schmitt, andR. H. Smith, Rare-eanh

distributions in themarine environment, J. Geophys. Res. 68, 4209-4217,

1963.

Haskin,M. A., andL. A. Haskin,Rareearthsin European shales: h redetermination,Science, 154, 507-509, 1966.

Hogdah!, O. T., S. Melson,andV. T. Bowen, Neutron activation analysis of elements in seawater,in Advances in Chemistry Series

competitive complexation of Y(iII) withrespect to thesurface , Vol.73, editedby R. A. Baker,308-325, American Chemical Society, 1968. organiccoatingsof marineparticlesand carbonate ionsin G., H. Elderfield, andA. Hudson,Rareearthelements in solutionis an importantfactorfor the fractionation between Klinkhammer, seawaternearhydrothermal vents, Nature 297, 185-188,1983. Y(I!I) and Ho. Our resultsof comparative Y/REE seawater of trivalentrareearth concentrationsis consistentwith the theory. Nevertheless, Lee, J. H., and R. H. Byrne, Complexargon elements(Ce, Eu, Gd, Tb, Yb) by carbonateions, Geochim, furtherstudyin neededto determine therelativeimportance of Cosraochim.Acta 57, 295-302, 1993. Y(III)/Ho(III) fractionargon duringparticle scavengingand Masuda, A., and Y. Ikeuchi, Lanthanidetetrad effectobserved in sediment diagenesis in controlling theirabundance in seawater marine environment, Geochem. J., 13, 19-22, 1979.

and marine deposits. Thus, the observation of Y(II!) canmake widerthe rangeof

Nance,W. B., andS. R. Taylor, Rareearthelementpatterns andcrustal evolution-I. Australianpost-Archean sedimentary rocks,Geochint. chemicalreactivityof REEs as an eka-lanthanide and should Cosraochirn.Acta, 40, 1534-1551 (1976)

help us betterunderstand the geochemistry of tracemetalsin Piepgras,D. I., and S. B. Jacobsen,The behaviorof rareea• elements in seawater: Precise. determination of variationsin the the ocean. Anotherinterestingelementin 3A groupof the North Pacific water column, Geochirn.Cosmochim. Acta, 56, periodictable is scandim which showssomewhatdifferent 1851-1862, 1992. chemistryfrom thoseof yttriumandREEs. Althoughthe study of Sc in the marineenvironmentis few, the Pacificprofile of Brewer et al. [1972] also shows the "nutrient-like" distribution.

The simultaneous measurements of Sc-Y-REEs

may further provide useful insights into the chemical systematicsin the oceans. Acknowledgments: We thank the Ministry of Education,Science

and Culture,Japanfor supportand Dr. R. Sherrellfor manuscript reading.

References

Rankin,P. C., and G. P. Glasby, Regionaldistributionof rareearth and minor elementsin manganesenodulesand associated sediments

in the southwest Pacificandotherlocalities, in Marine Geology and Oceanography of the Pacific ManganeseNodulesProvince, editedby J. L. Bischoff, and D. Z. Piper,681-697,Plenum Pub[, New York, 1979. Shabani,M. B., T. Akagi, H. Shimizu,and A. Masuda, Determination of tracelanthanides and yttriumin seawaterby inductively coupled

plasma mass spectrometryafter preconcentration with solvent extraction and back extraction, Anal. C hem., 62, 27092714,1990.. Sholkovitz,E. R., and D. L. Schneider, Cerium Redoxcycesandrare

earthelementsin the Sargasso Sea, Geochim.Cosmochirn. Acta, Bertram,C. I., and H. Elderfield,The geochemical ba!anceof the rare earthelementsand neodymiumisotopesin the oceans,Geochim. Coxmochirn.Acta 57, 1957-1986, 1993. Brewer, P. G., D. W. Spencer,and D, E. Robertson,Trace element

profilesfrom the GEOSECSII test stationin the Sargasso Sea, Earth Planet. Sci. Lett. 16, 11!-116, 1972.

Bruland,K. W., Oceanographic distributions of cadmium,zinc, nickel andcopperin thenorthPacific, EarthPlanet.Sci.Lett.,47, 176-

55, 2737-2743, 1991.

Sholkovitz,E. R., W. M. Landing, and B. L. Lewis, Oceanparticle chemistry:The fractionationof rare earth elementsbetween

suspended particles andseawater, Geochirn. Cosmochim. Acta58, 1567-1579, 1994.

Turekian, K. K., andK. H. Wedepohl,Distribution of theelements in somemajorunitsof theearthcrest,Bull.Geol.Soc.Am.72,175192, 1961.

198, 1980.

Byrne, R. H., and J. H. Lee, Comparativeyttrium and rare earth element chemistriesin seawater, Mar. Chem. 44, 121-130, 1993. DeBaar, H. J. W., M.P. Bacon, and P. G. Brewer, Rare-earth

distributionswith a positive Ce anomaly in the westernNorth Ariantic Ocean, Nature, 301,324-327, 1983. DeBaar, H. J. W., M.P. Bacon, P. G. Brewer, and K. Bruland,Rare earth elements in the Pacific and Atlantic oceans, Geochim. Cosmochim. Acta 49, 1943-1959, ! 985.

YoshiYuki N;'zaki, Hiroshi Amakawa, andJing Zhang, Ocean Research Institute,University of Tokyo,1-15-1Minamidai,

Nakanoku,Tokyo 164, Japan. (Email: [email protected])

(Received June30, 1994;RevisedSeptember 5, 1994; AcceptedSeptember 5, 1994)