Rep. Res. Edu. Ctr. Inlandwat. Enviren., 2 : 69--73 (2004)

Charactergzaag@n of fieedfivtidagag paptljcge$ gn water

$ag`gigpfies frdygye eke Fajfi agect $agaE"gigg Rfivers,

centrag Japan by SEMEDX

Tornohiro KYO"TANIi'2 and $ato$hi KOSHIMIZU"

i Earth Science Division, Yarfianashi InstitL}te of EnvironmeRtal Sciences, 5597-l, KeRrnarubi Kamiyoshida, Feg'iyoshida, YamaRashi 403-OO05, JapaR 2 Japan ScieRce and TechRology Corporation (JST), 4-l-8, }E[oftmachi, Kawaguchi, Saitama 332-OO12, Japan

ABSTRACT: The selective enrichment behaviors of labile substances sgch as hydroxides to the surface of specific mineral particles in river water were clarified by scaRniRg electron tn{croscopy/energy dispersive X-ray microanalysis (SEM-EDX). Individual particles with particle size between O.45 and 10pm other than diatom in water samples from the Fiiji and Sagami rivers, central Japan were aRalyzed by SEM-EDX aRd classified into seventeen groups accordii3g to the chemical compositioR and the shape. Phosphorus, sulfur, chlorine, maRganese and copper detected in each paytic}e could be used successfi;illy in order to elucidate the insoluble compoRents secondary formed iik river water, because the detection frequencies ofsuch elements are usually quite }ow or negligible iR fresh miReral particles derived fi"om igReous rocks. Almost all parts ofmanganese detected iR individual partic}es are ones adsorbed aftresh as the hydroxides together with iron and aluminum. Anions such as phosphorus, su}fiir and chlorine form the complexes with the hydroxides and/or are incorporated in them. The detection fi'eguencies of such iabiie species were predoininantly high iR Mg andfor Ca rich aluminosilicates.

Key Words: individual particle, river water, Mg andlor Ca rich alttminosilicates, scavenger of hydroxides, SEM-EDX

gntroduction

One of the process determining dissolved elemental concentration in river water is the interaction

between liquid and solid phases (Salbu and Steiftnes, 1995). However, the chemistry of fiver water has

been ever discussed concentrating on the dissolved species only (Salbu and Steinnes, l995; Koshimizu and

Kyotani, 2002). Consequently, the iRformation on the chemical compositioit ofsuspended particle in river water is not so much and the role in the elemektal fractionation between dissolved species and sgspended particle has not yet been clarified satisfactory. The main removal processes of dissolved elerReRts in

Ratural waters are coprecipltatlon witk hydroxides and the formation ofmetal-fumic cemplexes (Appelblad

et al., 1999; Haraguchi et al., 2000; Kawakubo et al., 2002; Kyotani and Koshimizu, 2000). ReceRtly, the

-69- latter has beeit intensively discussed from the speciation studies of trace e}ements by many researchers

(Appelblad et al., l999; Haraguchi et al., 2000; Kawakubo et al., 2002). While, although the fbrmer has been sometimes discgssed using labile iron existing in colloidal fraction from O.45pm fikrates (Kawakubo et al., 2002), the pardcles with size above O.45#m have been uswaIIy neglected.

The aim of this study is to clarify tke factors affecting the dissolved elemegta} concentrations in river water in terrns of tke chemical compositions of insoluble individual panicles. In this study, the chemical corripositions of iRdividL}al particles with size above O.45pm in river water were determined by scanniRg electroR microscopylenergy dispersive X-ray microanalysis (SEM-EDX). Phosphorus, sulfur, chlorine, maRganese and copper detected in individual particles were used in order to estimate the labile substances secondary forrned in river water. ConsequeRtly, tke selective enrichrneRt behaviors ofthe labile substances to specMc mineral particles were clarified.

Materials and Method$ Study area is the Fuji river whose geology is mainly andesitic and granitic, and the Sagami river characterized by mainly basaltic, central JapaR. Sampiing sites are showg in Fig.1. River vvater samples were collected in polypropylene coiktaiRer (l L) at each four poiRt in £he Faji (Fl-F4) and Sagami (Sl--S4) rivers on l Marcli, 2000. Tlie samples were shaken thoroughly prior to use. The 500 ml portion was imamediately filtered through a O.45ym PTFE membrane fiIter. The insoluble par£icies on the filter were detached by both sides-adhesive carbon tape and were fixed on tke aluminum sample stand for SEM-EDX. Thereafter, the particles were carbon coated (t"-200A) and tke specimens were subjected to SEM-EDX analysis. About 200 pardcles with particle size between O.45 and 10pm other than4s diatom iR each speci- men were ana}yzed under the condition N of accelerating voltage of 20kV; probe Mt.de Yatsugatake current of O.3 nA and counting time of To.kY IOO s. The characterlstic intensities of Shosenkyo the elements detected were normaiized Kofu to the total sum of characteristic X-ray basin Otsuki iRtensities of elements with atomic number over 1l to avoid tke particle MisakaMountains F2 Mma Sagami River size effect (Jambers and Van Grieken, Fl s2 s3 S4 F3 Sl 1997). Obvious slngle particles were F4 classified into sixteeR groups according A to the relative X-ray intensity. Ag- Mt, Fuji va muxws Shiraitonotaki gregates or biomineral-}ike substaRces Sagami Bay Fuji River were firstly discriiRinated from other gepugsstw particles by tke peculiar shape, and then IOkm were re-classified using the relative Suruga Bay

X-ray iRtensity. Fig. 1. Sampling locations.

-70- Re$ults and Discussion Particles adsorbing labile substances Fig.2 shows the abundaRces of phosphorus, sulfur, chlorine, inanganese and copper detected in individual particles from river waters and several source materials. Fig.2(a) expresses all particles other than diatom. As showR iR Fig.2(b), the detection flrequencies of these elements are usually quite low or negligible iR fresh mineral particles from igneous rocks, because tke detection limits of SEM-EI]}X are not enough to detect these elements being the contents be}ow approximately O.O1-O.!%. Thus, individual mineral panicies in rocks or atmospheric aerosol (Kyotani and Koshimizu, 2001) being the major source materials of suspended particles in river water do not contaiR significaRt}y above 5 elements. Therefbre, rriost ofthem detected in individual particies from river water can be considered to be not the constitueRts ofaluminosilicates but labile substances adsorbed afresh or secondary fbrmed in water. In the conven- tioRal SEM-EDX studies (Jambers and Van Grieken, l997), the contribution of }abiie species to individual pa4icle surfaces has not ever been considered sufficiently.

Also, the distribution iR river water of the insoluble manganese detected in iRdividual particles well corresponds to it of dilute-acid (O.e2M Ritric acid) soluble rnanganese measured in our previous work

(Kyotani and Koshimizu, 2000). These results suggest that most of rnanganese exists as easily sol"ble forms. It was a}so already showit that manganese is one of the predorninant elements together with iron and aluminum IOO in the dilute-acid (a) pa P III] S wa Cl [] Mn ee Cu soluble fractien from 80 ge suspended particles in o" 60 o = ff"., a river water and man- mp = ee = 40 ganese which can be D< 20 released easily from va -#- particles are mainly o # Fl F2 F3 F4 S2 hydroxides (Kyotani Sampling site and Iwatsuki, 100 (b) 1998; Kyotani and ee P E] S wa Cl [] Mn ee Cu 80 Koshimizu, 2000). * of 60 Thus the labile 9es , vsc manganese forras " 40 n< mainly the hydroxides 20 together with iabile o JG-la JG-2 JG-t3 i [IA-l [M-2 : JB-2 CJ-2 Oiivine iron and aluminum, : Granites I Andesites : Basalts (DH4908) : Chinese and adsorbs on the Granodiorites I : e Japaneseigneousrocks desert surfaces of mineral sand particles. At the same Fig. 2. Abundances of phosphorus, sulfur, chlorine, manganese and copper detected in time, it is suggested individual particles from river waters (a) and several seurce materials (b).

-71- that not only heavy meta}s but also anioRs suck as phospltorus, sulfur and chlorine form the coryiplexes

with the hydroxldes anCYor are iRcorporated in them. Judging from these points, the Fuji river has the kigh

removal eMciency of dissolved species.

Selective enrichreent of hydroxides to the sutface of specific minerals As a example, Fig.3 shows the abundance of the labile manganese by the particle type. The detection

frequeRcy of the labile manganese was predomiitantly high in Mg-Fe rich alumillosi}icates and aggregates

or biomiReral-like substances, especially in the Ftlji river. This result suggests that tke latter is aggregates

of hydroxides or humic substaBces. The fbnmer is a predomlnant component in solutions coRtaining the

iarge amounts of sol"bte components. However, tke former is an expedient existence, because almost all

parts of the manganese and most of the iroR can be regarded as secondary formed hydroxides, judgiRg from

their relative X-ray intensity ratio. Namely, Mg-Fe rich aluminosilicates classified expediently in Fig.3

shogld be considered to be essentially Mg and/or Ca rich aluminosilicates being the low iron contents.

Manganese was not detected significantly even in olivine being the typicai Mg-Fe rich aluminosilicates

as shown iR Fig.2(b). This result wel} corresponds to the feature of original Mg-Fe rich aiuminosilicates

from the Sagarni river wkich characterized by basaitic rocks. Nainely, original Mg-Fe rich alurfiinosilicates

do not adsorb significantly labile species. Furthermore, aggregates or biomiReral-like substances were re-

classified using the relative EOO X-ray intensity as shown classification by the relative intensity #･/ F2 .. / in Fig.3. As tlie results, 50 2e% se% ' she abttfidance of labile # o i manganese is changed ioo ' F3 *rf ew･ as foIlows. That is, it is classification by the relative intensity 50 1･,/,1'･,･,k. E 1･.･.･.･ 1OO% me ' predominan£ly high Rot b- ss o in Fe-rich particles but iB eq U-"I, o 1eo o v F4 Ca-rlch, and Mg andlor nc ca

/ttt.tt･//t･ vg 50 t./ttl･/t･t/ Ca rich aluminosilicates. S #wh < While, although Na-, o 100 K- and Na-Ca rich alumi- classification by the relative intensity 83% S2 nosilicates such as alka}i 50 l79x6 ttt･ feldspars or plagioclase ,twy- o xx sc whfi 6 are main rriinerals in the .9ts g :/, , ge, :. # # g･,L .9s ,9x o a2= x 'T" `Ta .crab ;;･ 8 8 Ee o U5 ra = .s ;xoth Fuji river, t}iey do not k E u . ･: ･: ･: w .9 signMcant}y adsorb the -i'n t 'tt A 8 6 2di 8re tz labile manganese. Tkus, twtKClyp: oa D alumiltesilicates 9 : the adsorption behaviors of co o Cb pa- Particle type at hydroxides to mineral par- Fig. 3. Abundance oflabile manganese by the pardcle type. ticles show the reraarkable *Mn-ricli particle was excluded from the calculation.

-72- selectivity. These results indicate tkat Mg andlor Ca rich aluminosilicates in river water play irr}portaitt roles as the scaveftgers of Mn-, Fe- and Al- hydroxides adsorbing dissolved trace elements aRd affect the elementa} fractionatioR between liquid and solid phases.

References

Appelblad, P.K., Baxter, D.C. and Tkunberg, J.O. (l999) DeterrninatioR ofmetal-liumic complexes,

free metal ions aRd total concentrations in natural waters. I Environ. Monit. I: 211--217. ' Haraguchi, H., Itoh, A., Ji, S. aiid Ylitbutani, T. (2000) Studies oR Bio-geochemical and eRvironmental

bekaviors oftrace eiements -chemical speciation oftrace elements in Ratgral water. Enviivn.

5tri., 13: 313-328.

Jambers, W. and Van Grieken, R. (1997) SiRgle partic}e characterization of iRorganic sgspension in

Lake Baikai, Siberia. Enviivn. Sei. 7iechnol., 31: l525-1533.

Kawakubo, S., 'fachikawa, K. and Iwatsuki, M. (2002) Speciation of iron in river water using a

specific catalytic determination and size firactioRation. X Environ. ?va)nit., 4: 263-269.

Koshimizu, S. and Kyotani, T. (2002) Geocheraical behaviors ofmultFelements in water samples from

the FuLji and Sagami rlvers, central Japan using vanadium as an effective indicator. lpn. Z

LimftoL 63, in press.

Kyotani, T. and Iwatsuki, M., (l998) Determination ofwater aRd acid soluble compoRents iR

atmosplkeric dust by inductively coupled plasma atomic emission spectrome£ry, ion chromatography and ion-selective electrode method. AnaL Sci., 14: 741-748.

Kyotani, T. and Koshimizu, S. (2000) Geochemical study of £race elemeRts in water samp}es from the

Fulji and Sagami rivers, cen£ral JapaR. 61 th Symp. Anal. Chem., abstract, l58. Kyotani, T. and Koshimizu, S. (200l) ldentification ofiRdividual Si-rick particles derived from Kosa

aerosol by the alkali elemental composition. Bull. Chem Soc. .ipn., 74: 723-729.

Salbu, B. and Steinnes, E. (eds.) (l995) Trace eiements in natural waters. CRC Press, LoRdon.

-73-