Nuclear Instruments and Methods in Physics Research B 181 (2001) 598±602

www.elsevier.com/locate/nimb

Characterisation of zinc in slags originated from a contaminated sediment by coupling l-PIXE, l-RBS, l-EXAFS and powder EXAFS spectroscopy M.P. Isaure b

a,b,*

, A. Laboudigue a, A. Manceau b, G. Sarret b, C. Ti€reau a, P. Trocellier c

a Centre National de Recherche sur les Sites et Sols Pollu es, BP537, 59505 Douai Cedex, France Environmental Geochemistry Group, LGIT-IRIGM, University of Grenoble, BP53, 38041 Grenoble Cedex 9, France c CEA-CNRS, Laboratoire Pierre Sue, CE Saclay, 91191 Gif-sur-Yvette Cedex, France

Abstract Depositing dredged sediments on soils is usual but it is a hazardous practice for the local environment when these sediments are polluted by heavy metals. This chemical hazard can be assessed by determining the speciation of metals. In this study, slags highly polluted with Zn and originated from a contaminated dredged sediment were investigated. Zn speciation was studied by laterally resolved techniques such as l-particle induced X-ray emission (l-PIXE), l-Rutherford backscattering spectrometry (l-RBS), l-extended X-ray absorption ®ne structure (l-EXAFS), and bulk analyses such as powder EXAFS spectroscopy. l-PIXE and l-RBS results showed that high concentrations of Zn were associated with S in localised areas at the surface of the slags while moderate amounts of Zn were mainly associated with Fe in the matrix. EXAFS results allowed to identify ZnS and Zn sorbed on ferrihydrite …5Fe2 O3
9H2 O†, proxy for iron oxy-hydroxides, as the main Zn-bearing phases. The occurrence of this Zn±iron oxy-hydroxide is interpreted as a mobilisation of Zn released from ZnS oxidation. Ó 2001 Elsevier Science B.V. All rights reserved. PACS: 89.60.+x; 82.80.yc; 61.10.Ht Keywords: l-RBS; l-PIXE; EXAFS; Zinc; Contaminated sediment

1. Introduction Sediments accumulating in ship-canals are periodically dredged for maintenance and are gen-

* Corresponding author. Tel.: +33-4-76-82-80-09; fax: +334-76-82-81-01. E-mail address: [email protected] (M.P. Isaure).

erally deposited along banks. However, in industrial areas, these sediments are often polluted with heavy metals, and these practices are hazardous because metals can migrate to the environment. This chemical hazard depends on the mobility and bioavailability of the metals and can be assessed by determining their speciation [1±3]. The aim of our study was to identify the Zn speciation and distribution in the coarse slags originated from a dredged contaminated sediment.

0168-583X/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 0 1 ) 0 0 5 2 9 - 8

M.P. Isaure et al. / Nucl. Instr. and Meth. in Phys. Res. B 181 (2001) 598±602

Due to the highly heterogeneous nature of the sediment, bulk analyses should be completed by investigations at the micrometer scale to determine Zn partitioning. l-particle induced X-ray emission (l-PIXE) and l-Rutherford backscattering spectrometry (l-RBS) are well adapted to our study because (i) l-PIXE gives information on lateral distribution of an element with Z > 12, (ii) l-RBS allows to obtain the distribution of the elements with depth, with a good selectivity for thin heavy element layers deposited on a light matrix [4], (iii) both techniques present multi-elemental analytical capabilities and high sensitivity. These techniques have been used successfully for the characterisation of environmental samples for several years [5± 8]. Bulk extended X-ray absorption ®ne structure (EXAFS) and l-EXAFS spectroscopy provide structural information on the studied element [3,9]. 2. Materials and methods 2.1. Samples The contaminated sediment (6600 ppm of Zn) was dredged from the Scarpe canal in the north of France, a region highly polluted by smelting activities. Immediately after dredging and depositing on a soil, the sediment was collected, air dried and fractionated by sieving and sedimentation at 2 mm, 500, 200, 50, 20 and 2 lm. Chemical analyses performed by ICP-AES on each grain size fraction showed that the coarse one (2000±500 lm) was highly polluted with Zn (8400 ppm). The observation with a binocular lens showed that this fraction was mainly composed of anthropogenic slags, related to surrounding smelting activities [10]. Among these slags, porous black slags (PBS) showed the highest Zn enrichment (11,000 ppm) and were investigated by l-RBS, l-PIXE, l-EXAFS and powder EXAFS. 2.2. Nuclear microprobe Analyses were performed with 3:07 MeV 4 He‡ using the Laboratoire Pierre S ue (LPS) nuclear microprobe facility in Saclay, France [11]. Using this beam allows a good mass resolution of RBS

599

results and, as it corresponds to the resonance energy of oxygen, enhances its signal. PIXE spectra were collected using a Si(Li) X-ray detector having a 7 mm diameter aperture. RBS spectra were collected with an annular surface detector of 100 mm2 active area and a thickness of 100 lm. A 18 lm Mylar ®lm was positioned in front of the Si(Li) detector to shield the detector from the backscattered 3:07 MeV 4 He‡ . PIXE analyses were calibrated with a PIXE standard containing Al, Ti, Cr, Co, Ni and Mo, and RBS analyses with a standard constituted of 20 nm Au deposited on Al. The complementary techniques l-PIXE and lRBS were used simultaneously to check the consistency of the results. For mapping, samples were prepared in 30 lm thick thin sections embedded in epoxy resin. 4 He‡ beam was focused to approximately 4 lm  4 lm: Elemental maps were obtained by scanning beam over areas of 155 lm  115 lm with an average current density of around a few pA/lm2 and an accumulated charge of approximately 2 lC. Point analyses were also performed on thin sections and on individual grains of PBS stuck in Ag lacquer on indium. These former analyses were collected with a current of approximately 100±200 pA and an accumulated charge ranging from 0.1 to 0:4 lC. l-PIXE spectra were analysed using the Gupix software [12] and RBS spectra using the SIMNRA software [13]. 2.3. EXAFS Powder EXAFS experiments were performed on the BM32 beamline at European Synchrotron Radiation Facility (ESRF) in Grenoble, France. Zn K-edge ¯uorescence yield powder EXAFS spectra of PBS pellets were recorded with a multielement Ge solid-state detector (Canberra). Zn-lEXAFS spectra were collected in ¯uorescence mode on the 10.3.2 station at the Advanced Light Source (ALS) in Berkeley, USA [14]. The beam was focused on 2 lm  2 lm. Samples were prepared in 30 lm thick thin sections embedded in epoxy resin. The speciation of Zn was determined using a linear combination of reference compounds in a least-square ®tting of unknown spectra. The frac-

600

M.P. Isaure et al. / Nucl. Instr. and Meth. in Phys. Res. B 181 (2001) 598±602

tional contribution of each reference spectrum to the ®t is directly proportional to the amount of Zn present in that form in the sample. A large database of reference Zn-containing minerals was generated in our laboratory including pure Zn minerals, Zn-substituted and Zn-sorbed references [3]. In the calculated spectra, the sum of percent-

age for Zn species was not always equal to 100% because a tolerance of 20% was admitted. Errors on the quanti®cation of Zn species were comprised between 10% and 20%.

3. Results and discussion l-PIXE elemental maps of a selected PBS grain showed a heterogeneous distribution of Zn, S and Fe (Fig. 1). The highest Zn concentrations were found in S-rich areas of several tens of micrometers. As an example, point analysis performed in point A indicated concentrations of 10,5 wt% for Zn and 8,6 wt% for S. l-PIXE data contained in Fig. 1 were used to draw the plot of Zn versus S lPIXE intensities (Fig. 2). This plot con®rms that Zn and S are strongly correlated (correlation coecient R ˆ 0:80) for the major proportion of points (group A). This group corresponds to (Zn, S)-enriched areas. However, Fig. 2 shows a second group of points (group B) in which Zn and S are less correlated. Indeed, a di€use association between moderate concentrations of Zn and Fe could be observed in the elemental maps. Point analyses performed in these areas con®rmed moderate concentrations of Zn and Fe, and the low S concentration. For instance, in point B (Fig. 1), 14300 ppm of Zn, 48000 ppm of Fe and 320 ppm of S were measured. l-RBS analysis performed on individual grains of PBS showed two major peaks corresponding to Zn and S, indicating that their distribution changes with depth (Fig. 3(a)). Calculation performed

Fig. 1. l-PIXE elemental maps showing the distribution of Zn, S and Fe in a PBS (scanning step: 20 lm, count time: 100 ms/ pt). Concentrations measured in point A are: [Zn] 105000 ppm, [S] 86000 ppm and [Fe] 47000 ppm. Concentrations measured in point B are: [Zn] 14300 ppm, [S] 320 ppm and [Fe] 48000 ppm.

Fig. 2. Plot of l-PIXE intensities showing the pattern of Zn±S relationship in the PBS grain shown in Fig. 1.

M.P. Isaure et al. / Nucl. Instr. and Meth. in Phys. Res. B 181 (2001) 598±602

with SIMNRA showed that the experimental spectrum could be ®tted by assuming the presence of three layers, with decreasing Zn and S concentrations from the surface to the interior. The surface layer has the following composition, in atomic percentage: S 38%, Zn 29%, O 10%, Fe 7%, Ti 6%, Ca 5%, C 4%, Cd < 1%, Pb < 1%: Considering low amount of oxygen in this layer, a reduced (Zn, S) phase is suggested. In the intermediate layer, Zn and S concentrations are lower (10% and 15%) while Si, O, C, Ca, P and Fe constitute the major layer components. The deep layer is mainly composed of Si, O, Ca, Mg, Fe. l-PIXE spectrum collected in the same point of analysis con®rmed the dominance of S and Zn in this area (Fig. 3(b)). l-EXAFS spectrum collected in the (Zn, S)enriched area of the slag could be reproduced by 108%  20% ZnS (sphalerite), indicating that sphalerite is the only Zn-bearing phase in this area (Fig. 4(a)). EXAFS spectrum recorded on the

Fig. 3. l-RBS spectrum and its simulation (a), and l-PIXE spectrum (b) …3:07 MeV 4 He‡ †, for a PBS grain.

601

powder of PBS was not correctly ®tted by considering only one Zn species. Adding a second Zn component improved the ®t and the experimental spectrum was correctly ®tted with a mixture of 82% ZnS ‡ 36% Zn sorbed on ferrihydrite (Fig. 4(b)). These EXAFS results are consistent with the nuclear microprobe results which indicated a Zn/S association in a reduced form on localised area and a Zn/Fe association in a more di€use area. ZnS was thus identi®ed as the main Zn component in PBS and is likely to originate from anthropogenic pollution. Indeed, ZnS was identi®ed in dusts originating from smelting activities in the same region [15]. Zn sorbed on ferrihydrite, proxy for iron oxy-hydroxides, was the minor Zn phase

Fig. 4. l-EXAFS spectrum and its simulation for the (Zn, S)enriched area in PBS grain (a), and powder EXAFS spectrum and its simulations for the PBS powder (b).

602

M.P. Isaure et al. / Nucl. Instr. and Meth. in Phys. Res. B 181 (2001) 598±602

identi®ed in PBS. This poorly crystallised iron oxide is relatively common in surface environment and easily transforms into more stable iron oxides as goethite or hematite [16]. Therefore, Zn-containing ferrihydrite formation is likely to be the youngest phase in PBS and its occurrence can be interpreted by the oxidation of ZnS and FeS2 (identi®ed elsewhere in the sediment). 4. Conclusion In the studied PBS originated from a contaminated sediment, Zn is predominantly present as anthropogenic ZnS and, to a lesser extent, as Zn± iron oxy-hydroxide. ZnS is distributed on small areas at the surface of the grains whereas Zn±iron oxy-hydroxide is more di€use. This secondary Zn± Fe species results from the di€usion in PBS of dissolved Zn released by the weathering of ZnS. Owing to the heterogeneity of natural samples, combined use of laterally resolved techniques such as l-PIXE, l-RBS and l-EXAFS, and bulk techniques such as EXAFS proved complementary in the investigation of Zn partitioning and speciation. Acknowledgements The authors acknowledge J.L. Hazemann and O. Proux on the BM32 beamline at European Synchrotron Radiation Facility (ESRF) in Grenoble, and G. Lamble and R. Celestre on the 10.3.2 beamline at the Advanced Light Source (ALS) in Berkeley, for their assistance in the collection of EXAFS and l-EXAFS spectra, respectively. We also acknowledge L. Daudin and all sta€ of the Laboratoire Pierre S ue (LPS) in Saclay for their assistance in collecting nuclear microprobe results and their helpful advice. We are

grateful to the ESRF, the ALS and the LPS for the provision of beamtime.

References [1] A. Manceau, M.C. Boisset, G. Sarret, J.L. Hazemann, M. Mench, P. Cambier, R. Prost, Environ. Sci. Technol. 30 (1996) 1540. [2] G. Morin, J. Ostergren, F. Juillot, P. Ildefonse, G. Calas, G.E. Brown Jr., Am. Mineral. 84 (1999) 420. [3] A. Manceau, B. Lanson, M.L. Schlegel, J.C. Harge, M. Musso, L. Eybert-Berard, J.L. Hazemann, D. Chateigner, G. Lamble, Am. J. Sci. 300 (2000) 289. [4] G. Revel, Ann. Chim. Fr. 19 (1994) 209. [5] P. Cruvinel, R. Flocchini, P. Artaxo, S. Crestana, P.S.P. Herrmann Jr., Nucl. Instr. and Meth. B 150 (1999) 478. [6] J. Mesjasz-Przybylowicz, K. Grodzinska, W. Przybylowicz, B. Godzik, G.S. Lukaszewska, Nucl. Instr. and Meth. B 158 (1999) 306. [7] I. Orlic, S. Tang, Nucl. Instr. and Meth. B 150 (1999) 291. [8] P. Trocellier, V. Badillo, N. Barre, L. Bois, C. Cachoir, J.P. Gallien, S. Guilbert, F. Mercier, C. Ti€reau, Nucl. Instr. and Meth. B 158 (1999) 511. [9] B.K. Teo, EXAFS: Basic Principles and Data Analysis, Springer, Berlin, 1986, p. 349. [10] M.P. Isaure, A. Laboudigue, A. Manceau, G. Sarret, C. Ti€reau, P. Lecomte, in: Proc. of the Seventh International Conference on Contaminated Soil, Leipzig, Germany, 18± 22 September 2000, p. 734. [11] P. Trocellier, Microsc. Microanal. Microstruct. 7 (1996) 235. [12] J.A. Maxwell, J.L. Campbell, W.J. Teesdale, Nucl. Instr. and Meth. B 43 (1989) 218. [13] M. Mayer, SIMNRA User's Guide, Tech. Rep. IPP 9/113, Max-Planck-Institut f ur Plasmaphysik, Garching, 1997. [14] A. MacDowell, C. Chang, G. Lamble, R. Celestre, J. Patel, H. Padmore, in: Proceedings of the SPIE Conference on Xray Microfocusing: Applications and Techniques, Vol. 3449, San Diego, California, 1998, p. 137. [15] S. Sobanska, N. Ricq, A. Laboudigue, R. Guillermo, C. Bremard, J. Laureyns, J. Merlin, J. Wignacourt, Environ. Sci. Technol. 33 (1999) 1334. [16] R.M. Cornell, U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurrence and Uses, VCH Publishers, New York, 1996, p. 573.