Elimination of interferences in determination of platinum and palladium in environmental samples by inductively coupled plasma mass spectrometry

Analytica Chimica Acta 564 (2006) 236–242

Elimination of interferences in determination of platinum and palladium in environmental samples by inductively coupled plasma mass spectrometry a , Anna Ruszczy´ ˙ Barbara A. Le´sniewska a,∗ , Beata Godlewska-Zyłkiewicz nska b , Ewa Bulska b , Adam Hulanicki b a

Institute of Chemistry, University of Białystok, Hurtowa 1, 15-399 Białystok, Poland b Faculty of Chemistry, Warsaw University, Pasteura 1, 02-093 Warsaw, Poland

Received 20 October 2005; received in revised form 16 January 2006; accepted 19 January 2006 Available online 9 March 2006

Abstract Various approaches were evaluated in order to eliminate the spectral interferences noted when Pt and Pd has to be determined in environmental dust samples by ICP-MS. The chemical separation of Pt and Pd from the matrix components on ion-exchange resins was applied. The performance of cation-exchange resins (Dowex 50 WX-8, Dowex 50 WX-2, Dowex HCR-S, Varion KS, Cellex-P) for the separation of interfering ions was then examined. It was found that Dowex 50 WX-8 shows best performance. The effects of mass, mesh number of resin and concentration of Cl− ions on matrix separation were also studied. Another approach was to use the anion-exchange sorbent Cellex-T, which allows almost total retention of both analytes followed by their elution with 0.1 mol L−1 thiourea in 1 mol L−1 HCl. This procedure however can be used only for platinum determination by ICP-MS. The accuracy of proposed procedures was confirmed by the analysis of certified material BCR-723, and then it was used for determination Pt and Pd in samples of road dust. © 2006 Elsevier B.V. All rights reserved. Keywords: Cation-exchanger; Anion-exchanger; Sample pre-treatment; Matrix separation; Road and tunnel dust

1. Introduction The introduction of catalytic converters of automobiles containing platinum, palladium and rhodium (platinum group elements, PGEs) for reducing emission of gaseous pollutants, such as carbon monoxide, nitrogen oxides and hydrocarbons, has resulted in increasing concentration of PGEs in environmental matrices, especially in roadside dust, soil and plants [1,2]. The presence of Pt and Pd in such matrices can effect the human health and living organisms [3,4], therefore the investigation and control of PGEs content in environment are necessary. The determination of Pt and Pd in complex matrices, such as environmental or biological samples is still a difficult task due to their extremely low concentrations and significant matrix effects. Most sensitive analytical methods, as graphite furnace atomic absorption spectrometry (GFAAS) [5–8], instrumental

∗

Corresponding author. Tel.: +48 85 7457806; fax: +48 85 7470113. E-mail address: [email protected] (B.A. Le´sniewska).

0003-2670/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2006.01.066

neutron activation analysis (INAA) [9], adsorptive stripping voltammetry (ASV) [10], inductively coupled plasma-optical emission spectrometry (ICP-OES) [11] and inductively coupled plasma-mass spectrometry (ICP-MS) [12–28] are commonly used for that purpose. Due to the very low detection limit and multielement capability of ICP-MS, this method is considered as the most appropriate for direct determination of PGEs in various environmental samples. Unfortunately, complex matrices of environmental samples are often a source of spectral and nonspectral interferences caused by monoatomic and polyatomic ions, formed in the plasma from the matrix constituents. Overlapping of interferent and analyte signals hampers obtaining accurate results. The contribution of interferent in analyte signal depends on the concentration of both, the analyte and interferent as well as the ability of the formation of interfering species in the plasma. Mathematical correction, based on the evaluation of the contribution of interferent signal in that one for analyte, was successfully applied for elimination of spectral interferences occurring during determination of Pt by ICP-MS tech-

B.A. Le´sniewska et al. / Analytica Chimica Acta 564 (2006) 236–242

nique [12–16], but was inefficient during determination of Pd [12,14–16]. Hence, the accurate results may be obtained only after chemical separation of Pt and Pd from the matrix. Among different separation techniques, solid phase extraction based on adsorption [17], ion-exchange [18–27] or chelating properties [28] of solid sorbents have been demonstrated to be most effective. The ion-exchange methods exploit high affinity of Pt and Pd chlorocomplexes for anion-exchange resins and their low affinity for cation-exchange resins. The effective separation of Pt and Pd from the matrix of large amounts of base metals requires using large cation-exchange columns [18–21]. The other disadvantage of such procedures is partial retention of Pt and Pd on the column [29]. Because of low concentration of Pt and Pd in environmental samples, the application of anion-exchange resins seems to be preferred over the cation-exchange resins. However, the main limitation in using strong anion-exchangers is a non-selective retention of Pt and Pd chlorocomplexes on commercially available resins. Other elements that form anionic complexes (i.e. Al, Fe, Pb, Zn, Ni, Cu, Y) can be retained simultaneously with Pt and Pd, introducing spectral interference during detection step [11,22,30]. In some cases Pt and Pd are so strongly bound to the resin that for their quantitative recovery the use of high volumes (40–75 mL) of concentrated mineral acids [22–26] or total digestion of the resin [28,30] were proposed. In addition, when hot nitric acid was used for elution of Pd and Pt from AG 1-X8 resin the unexpected interferences were observed, probably due to partial digestion of the resin [22]. Some authors reported that in order to achieve accurate results for determination of Pt and Pd by ICP-MS, the combination of cation or anion exchange methods with isotope dilution technique (ID) was necessary [20,21,24,25]. To overcome these limitations we have evaluated and critically compared procedures based on different cation-exchanger and cellulose anion-exchanger for elimination of interferences during determination of Pt and Pd by ICP-MS. Cellulose sorbents are characterised by lower affinity to platinum group metals [5,6,31], what facilitates the elution process and in consequence assures the quantitative recovery of retained metals with small volume of eluents. 2. Experimental 2.1. Instrumentation A quadrupole spectrometer (ICP-QMS) (ELAN 6100 DRC, Perkin-Elmer Sciex, USA) equipped with conventional Scott spray chamber and Meinhard nebulizer was used in this study. The working conditions of spectrometer were optimized daily in order to obtain the maximal sensitivity and stability as well as the lowest level of oxides and double charged ions. The operating conditions are given in Table 1. The isotopes 105 Pd and 195 Pt were used for quantification. The signals of other isotopes of analytes (104 Pd, 108 Pd, 194 Pt, 196 Pt) and interfering elements (65 Cu, 89 Y, 87 Sr, 179 Hf) were also monitored. A PU 9100X (Philips Scientific, Cambridge, UK) atomic absorption spectrometer equipped with a PU 9390X electrother-

237

Table 1 ICP-MS (ELAN 6100 DRC) operating conditions Plasma RF power Frequency Nebulizer gas flow rate (Ar) Plasma gas flow rate (Ar) Auxiliary gas flow rate (Ar) Data acquisition Mode Dwell time Number of sweeps per reading Number of replicates Analytical masses Interfering ions Internal standard

1100 W 40 MHz 0.93 mL min−1 15.5 L min−1 1.0 L min−1 Electric scan 100 ms 6 3 105 Pd, 195 Pt 105 Pd: 40 Ar65 Cu+ , 89 Y16 O+ , 88 Sr16 OH+ , 88 Sr17 O+ ; 195 Pt: 179 Hf16 O+ 115 In

mal atomizer, FS-90 autosampler and deuterium background corrector was used for Pd and Pt determination during optimization of the separation procedures. The palladium and platinum hollow cathode lamps (CPI, USA) were operated at 8 and 10 mA current, respectively. The absorbance was measured in the peak height mode using pyrolytically coated graphite furnaces with 0.5 nm spectral bandpass at 247.6 nm for Pd and 265.9 nm for Pt. Sample aliquots of 20 ␮L volume were dispensed from autosampler cups into the graphite furnace. The following optimized time/temperature furnace programs for Pd and Pt determination were used: drying: 110 ◦ C for 10 s, 160 ◦ C for 10 s, ashing: 350 ◦ C for 10 s, 1300 ◦ C (Pd) or 1600 ◦ C (Pt) for 5 s, atomization: 2400 ◦ C (Pd) or 2700 ◦ C (Pt) for 2.5 s. Microwave system ETHOS PLUS (Milestone, Italy) was used for digestion of samples. A solid-phase extraction system SPE-12G (Baker, Germany) equipped with cation-exchange columns and a system consisted of peristaltic pump Minipuls 3 (Gilson, France), PTFE tubing of i.d. 0.8 mm and anion-exchange columns were used for separation. 2.2. Reagents and materials Concentrated HNO3 (69%), HCl (37%) and HF (40%) (Suprapur, Merck) were used throughout this study. High-purity deionized water for the preparation of standards and sample solutions was obtained from Milli-Q system (Millipore, USA). Working solutions were prepared daily by dilution of stock solutions of 1000 mg L−1 of Pd and Pt (SPEX, Edison, NJ, USA) with high-purity water. Single-element stock solutions of Cu, Sr, Y, Hf, and In as internal standard (1000 mg L−1 , SPEX, Edison, NJ, USA) were used. Thiourea (puriss p.a., Fluka, Germany) was used for elution of Pt and Pd from anion-exchange resin. Cation-exchange resins: Dowex 50 WX-8 (Fluka, Germany), Dowex 50 WX-2, Dowex HCR-S (polystyrene matrix, sulfonic acid functional group; Dow Chemical Co, USA), Varion KS (styrene–divinylobenzene matrix, sulfonic acid functional group; Nitrokemia, Hungary) and Cellex-P (cellulose matrix, phosphonic acid functional group; Bio-Rad Laboratories, USA)

238

B.A. Le´sniewska et al. / Analytica Chimica Acta 564 (2006) 236–242

in H+ form and anion-exchange sorbent Cellex-T, highly purified cellulose powder containing quaternary amine exchange groups (triethylaminocellulose, TEAE, –C2 H4 N+ (C2 H5 )3 , BioRad Laboratories, USA) were used for separation. The certified reference material of road dust BCR-723 (IRMM, Geel, Belgium) and candidate for reference material CW-7 (European Commission, Joint Research Centre, ISPRA) were used for method validation and to established the accuracy of the investigated procedures. The samples of road dust and tunnel dust collected in the city of Białystok in years 2000–2003, according to the procedure described in [12], were used in this study. 2.3. Procedures 2.3.1. Sample digestion procedure An amount of 100–150 mg of dust was weighted in highpressure Teflon vessels, to which 6 mL of conc. HCl and 2 mL of conc. HNO3 were added and the vessels were left loosely capped for 5 min. Next, the vessels were closed and samples were heated according to optimized microwave program: 250 W for 5 min, 450 W for 10 min and 650 W for 15 min, which was repeated three to four times to obtain the total dissolution of Pd and Pt. Alternatively the procedure including addition of HF [12] was performed. After digestion, samples were cooled to ambient temperature and vented carefully. The obtained solutions were evaporated almost to dryness in quartz crucibles. The residues were dissolved in 2 mL of conc. HCl and heated again almost to dryness. This treatment was repeated twice to reach complete dissolution. The final residues were dissolved in 14 mL of 0.5 mol L−1 HCl. 2.3.2. Separation of matrix ions on ion-exchange resins All preliminary examinations of resins performance were performed with the use of GFAAS on model solutions containing Pt and Pd and potential interferents, and spiked samples. In all cases, both elements were considered as to be assured, that after treatment of the samples, Pt and Pd could be determined simultaneously by ICP-MS. The final optimisation for real samples was performed with the use of ICP-MS. Before use, each cation-exchange resin (20 g) was carefully cleaned by stirring with 20 mL of 4 mol L−1 HCl solution (six times), next with Milli-Q water (twice) and 0.5 mol L−1 HCl (twice) for about 20 min. The laboratory-made columns (polyethylene tube, 80 mm × 8 mm i.d.) were packed with 3.5 g of the clean resin. Both ends of column tube were blocked with PTFE membranes. The resin was equilibrated with 10 mL of 0.5 mol L−1 HCl. The blank, obtained by passing 0.5 mol L−1 HCl through the column, was always determined prior to loading the sample in order to control possible contamination. Each column with low blank (<100 cps for Pt and Pd) was used once for matrix separation. In order to separate interfering elements 5 mL of sample solution prepared in 0.5 mol L−1 HCl was passed through the precleaned resin at the flow rate of 1 mL min−1 . The effluent containing Pd and Pt was collected. Next, the resin was rinsed with 2 mL of 0.5 mol L−1 HCl and this solution was also

added to the effluent, in which Pd and Pt were determined by GFAAS or ICP-MS, depending on the concentration of analytes. The concentration of interfering elements in effluent was also monitored. The separation of Pd and Pt as chlorocomplexes was carried out on the anion-exchange sorbent Cellex-T. The laboratorymade columns (polyethylene tube, 50 mm × 4 mm i.d.) were packed with 0.2 g of Cellex-T, initially conditioned with 5 mL of Milli-Q water or 5 mL of 0.25 mol L−1 HCl. For the retention of Pd and Pt, 5 mL of sample solution of pH about 1 (when column was conditioned with HCl) or 4 (when column was conditioned with water) was passed through the column at the flow rate of 3 mL min−1 . Next the column was rinsed with 2 mL of 0.1 mol L−1 HNO3 in order to remove retained interfering ions. The both analytes were eluted simultaneously by pumping of 4 mL of eluent solution (0.1 mol L−1 thiourea in 1 mol L−1 HCl) at the flow rate of 0.3 mL min−1 . The concentration of Pd and Pt in obtained solution was determined by GFAAS or ICP-MS. The concentration of interfering elements was also monitored. 2.4. Quantification of analytes The calibration graphs for determination of Pd and Pt by ICP-MS were prepared in the range from 10 to 400 ng L−1 . All standard solutions were prepared either in 0.2 mol L−1 HCl (for samples after cation-exchange separation) or in 0.02 mol L−1 thiourea in 0.2 mol L−1 HCl (for samples after anion-exchange separation). Determination of Pt and Pd in dust samples by mathematical correction method was performed according to the procedure described in [12]. All samples and standards were spiked with 2 ␮g L−1 of indium as an internal standard to compensate any changes of the signal due to the matrix. The calibration graphs for determination of Pd and Pt by GFAAS were prepared in the range from 5 to 100 ␮g L−1 for Pd and from 5 to 150 ␮g L−1 for Pt. In all cases, the standards of Pd or Pt were prepared in eluent solution. 3. Results and discussion The separation of Pd and Pt from the interfering ions on ion exchange resins is possible due to the ability of these metals to form highly stable anionic complexes in hydrochloric acid solution, whereas most transition metals, alkali and alkalineearth metals form weaker anionic or stable cationic complexes. Different factors, such as the form of analyte and its charge, acidity of the sample, presence of other ions in the solution as well as the kind of used resin can influence the separation process. The chemistry of Pd and Pt solution is complex with domination of coordination compounds of different stability and kinetic properties. Various complexes of Pd(II), Pt(II) and Pt(IV) can be formed in aqueous solutions depending on the concentration of chloride ions and the acidity of sample. Pd and Pt have high tendency to undergo hydrolysis. In dilute stock solutions of PtCl4 2− and PtCl6 2− , apart from mother ions, at least three Pt(II) and five Pt(IV) hydrolytic species of different charge (e.g. PtCl5 (H2 O)− and PtCl3 (H2 O)− ) are formed within several

B.A. Le´sniewska et al. / Analytica Chimica Acta 564 (2006) 236–242

hours [32]. In acidic solutions of Pd (pH 1–3) in the presence of 0.01 mol L−1 Cl− ions the predominant Pd species are PdCl4 2− , PdCl3 (H2 O)− and PdCl2 (H2 O)2 [33], while in the presence of 0.1 mol L−1 Cl− ions the main form of Pd (∼95%) is PdCl4 2− [34]. At pH above 3 Pd(OH)2 and Pd(OH)4 2− may precipitate. When concentration of Cl− ions is lower than 0.001 mol L−1 , the cationic Pd species, such as PdCl(H2 O)3 + , Pd2+ , Pd(OH)+ are predominant [33]. At pH < 0.5 the exchange efficiency becomes poorer because the partition ratio of interfering cation between the sample solution and resin is low. Additionally, in such solutions the anionic complexes of matrix ions (e.g. FeCl4 − , HfF6 2− ) can be formed which are not retained on cation-exchange resin. At lower acidity, the H+ of such resin can be easier replaced by the interfering cation. Therefore, during optimization of sample pre-treatment step special attention was paid to the presence of analytes as well as matrix ions in the forms suitable for their further separation. 3.1. Performance of cation-exchange resin It was expected, that at acidic media the cation-exchanger would serve as an absorber for interfering ions, while both Pt and Pd should stay in the solutions and therefore could be determined directly in the effluent. The resins with various polymeric matrix and functional groups as: Dowex 50 WX-2, Dowex 50 WX8, Dowex HCR-S, Varion KS and Cellex-P were tested. Both, the recovery of Pd and Pt and the retention efficiency of the potentially interfering metals (Cu, Ni, Co, Fe, Pb) from solutions containing analytes and interferents were assessed. As was discussed above the low pH and the presence of appropriate concentration of Cl− ions in the solution are necessary to provide the occurrence of Pd and Pt in anionic form. Therefore in the first instant, the influence of different concentration of HCl (0.03–0.5 mol L−1 ) on the efficiency of retention of interfering cations on studied resins was investigated. It was found that the total retention of Fe(III) ions occurred from its 0.5 mol L−1 HCl solution, whereas Su et al. [35] reported that 0.05 mol L−1 HCl solution is sufficient to separate cations of matrix from Pt and Pd on strong cation-exchanger (type 723). The capacity of resins for matrix ions was tested by passing through the column increasing masses (up to 4000 ␮g) of Fe(III) ions. It was found that in the presence of 0.5 mol L−1 HCl up to 500 ␮g of Fe(III) ions are totally retained on Dowex 50 WX-2 and Dowex 50 WX-8 resins. The slightly worse efficiency was noted for Varion KS (93%) and Dowex HCR-S (81%), and only 28% of Fe(III) was retained on Cellex-P. It should be stressed that with the applied conditions about 5% of Pt and Pd were simultaneously retained on the column. To overcome this effect, we proposed to introduce the additional step of rinsing out of both analytes from the column by passing of 2 mL of 0.5 mol L−1 HCl. This solution was always added to the first effluent. With this approach total recovery (100–104%) of Pt (75 ␮g) after its separation from 500 ␮g of Fe(III) ions was achieved for all tested sorbents. In the case of Pd (50 ␮g) its recovery was only about 94% after matrix separation on Dowex HCR-S and Cellex-P, whereas on other sorbents was satisfactory: 100.7 ± 2.3% on Dowex 50 WX-2, 104.0 ± 2.4% on Dowex 50 WX-8 and 100.0 ± 1.3% on Varion

239

Fig. 1. Recovery (±S.D.%, n = 5) of Pt (250 ng), Pd (150 ng) and Pb (1 mg), Ni (0.5 mg), Co (0.5 mg), Fe (1 mg) on cation-exchangers determined by GFAAS: (A) Dowex 50 WX-8, 20–50 mesh, 2g; (B) Dowex 50 WX-8, 50–100 mesh, 2 g; (C) Dowex 50 WX-8, 50–100 mesh, 4 g; (D) Varion KS, 4 g.

KS. Further studies were conducted with Dowex 50 WX-8 and Varion KS. The effects of mass of the Dowex 50WX-8 resin and its mesh number on the separation of matrix ions were also studied. In that case the solutions containing 250 ng of Pt, 150 ng of Pd and 1 mg of Zn(II), Pb(II), Fe(III), 0.5 mg of Ni(II), Co(II) and 0.25 mg of Cd(II) were passed through the resin. The concentration of all listed elements in effluents was determined by GFAAS. It was found that the efficiency of sorption of interfering ions is higher when the mesh of the resin is greater, as well as its mass is bigger (Fig. 1A–C). Probably, this is due to reduced flow rate of sample through the column and consequently longer contact time between ions and the surface of resin. The performance of Varion KS was found to be as good as of Dowex 50 WX-8 (Fig. 1D), but the number of counts measured by ICP-MS for the blank after passing the acidic solution through the columns filled with Varion KS and Dowex 50WX-8 was much higher for the former one. Then, we decided to use Dowex 50WX-8 for further investigations. In order to evaluate the performance of Dowex 50WX-8 for determination of Pt and Pd by ICP-MS the isotopic ratios of most abundant isotopes of Pd (104 Pd, 105 Pd, 108 Pd) and Pt (194 Pt, 195 Pt, 196 Pt) were measured in effluate and it was found to be as the naturally occurred ratio. The obtained recovery of analytes (1 ng Pd, 1 ng Pt) from interfering elements (0.05–25 ␮g of Hf, Sr, Y, Cu) in solution prepared in 0.5 mol L−1 HCl was 100.2 ± 1.8% (n = 6) for Pt and 90.9 ± 2.8% (n = 6) for Pd (Fig. 2A). Besides the use of model solutions (containing standards and studied interferents), it was of interest to evaluate the performance of the proposed approach for the real environmental dust samples after pre-treatment procedure. In this case samples were digested with different mixtures of acids (HCl, HNO3 , HF). It was found that presence of HF in the sample solution diminishes the efficiency of retention of the Y and Hf on cation-exchanger (Fig. 2B and C). Depending on the sample type (tunnel dust, road dust) only 51–75% of Y and 90–95% of Hf were removed from the solution. This could be explained by the formation of neutral or anionic fluorocomplexes of yttrium and hafnium. The more detailed discussion of this problem is given in our previous work [36]. The high concentration of Y and Hf in effluent solution con-

240

B.A. Le´sniewska et al. / Analytica Chimica Acta 564 (2006) 236–242

Fig. 2. Recovery (±S.D.%, n = 5) of Pt and Pd after matrix separation on Dowex 50 WX-8 (3.5 g, 200–400 mesh) determined by ICP-MS from: (A) model solution (1 ng Pt, 1 ng Pd, 0.5–25 ␮g Hf, Zr, Y, Sr, Cu); (B) tunnel dust digested with aqua regia + HF; (C) road dust digested with aqua regia + HF; (D) road dust leached with aqua regia; (E) BCR-723 leached with aqua regia. (B), (C), (D) data were compared to Pd and Pt content obtained by TXRF and ICP-MS after mathematical correction, respectively.

taining studied analytes results in incorrect recovery of Pt and Pd exceeding 110 and 200%, respectively. Leaching of samples only with aqua regia provided the suitable form of analytes and interfering elements for their separation on cation-exchange resin and satisfactory recovery of both elements (Fig. 2D and E). In order to evaluate the accuracy of the selected procedure for simultaneous determination of Pt and Pd in environmental dust samples by ICP-MS technique, the analysis of BCR-723 and CW-7 was performed. From the obtained results (Table 2) it is evident that separation of interfering ions on Dowex 50WX8 from samples leached with aqua regia was effective, without losses of analytes and introduction of any contamination. Therefore, the proposed procedure could replace the expensive Isotopic Dilution (ID) approach [21] as well as the use of complicated algorithms developed for calculating of analytes concentration, as it was proposed by other workers [20]. The applied separation unit utilising commercially available resin is simple, inexpensive and easily accessible in routine laboratories. 3.2. Performance of anion-exchange resin It is assumed that after digestion of dust samples in aqua regia and evaporation with HCl, both Pt and Pd are present in a form of

anionic chlorocomplexes and as such could be retained on anionexchange resins [22–25]. The separation of PGEs was most often performed on strong anion-exchangers based on polystyrene matrix [22–25,30]. As it was already mentioned, obtaining the total recovery of analytes from such resins implies using of concentrated acids [22–25] or digestion of resin [30]. The application of sorbents characterised by lower affinity for noble metals, e.g. alumina (Al2 O3 ) [38] or cellulose resins [5,6,31] can facilitate the elution process. Cellulose resin CellexT was used for the separation of Pd from Ir or Pt during analysis of alloys [39] and ores [40]. The elution of Pd from this resin was performed with 80 mL of 0.01 mol L−1 glycine solution [39] or 20 mL of 2 mol L−1 HNO3 [40]. In our former works [5,6] we have also used Cellex-T for preconcentration of Pt and Pd, and both analytes were eluted from the column with thiourea solution. Application of stronger eluent resulted in lower volume of reagent necessary for quantitative recovery of both analytes (0.3 mL of 1.2 mol L−1 thiourea (pH ∼0.5)) before their determination by GFAAS. Cellulose sorbent functionalised with 2,2 diammino-diethylamine (DEN) was used for separation of Pt(II) from medicines, however the initial reduction of Pt(IV) to Pt(II) was necessary [41]. In this case the authors have obtained the good recovery of the analyte with 0.2 mL of 0.1 mol L−1 thiourea in 0.5 mol L−1 HCl solution. In other works, when the separation of noble metals was performed on Dowex 1-X8 [23] or Dowex 1-X10 [11] polystyrene resins, the efficiency of removing of analytes from the resin with thiourea was low and high volumes of eluent solution (40–75 mL) had to be used. Jarvis et al. [23] have found that the additional step of elution with concentrated hydrochloric acid was necessary for quantitative recovery of PGEs. They have also observed some problems with using this reagent for ICP-MS determination because of a high level of total dissolved solids (TDS). In order to reduce the volume of thiourea, Kovacheva and Djingova [11] have proposed the elution of Pt and Pd at higher temperature as well as its 10-fold circulation through the column. Considering mentioned above problems the previously proposed procedure [6] was adapted for Pt and Pd determination in dust samples by ICP-MS. For this purpose the ability to use less concentrated solution for performing the elution of Pt and Pd from the anionic resin was tested. The digested dust samples

Table 2 The content of Pt and Pd in BCR-723, CW-7 and dust samples determined by ICP-MS after leaching of analytes with aqua regia using different procedures of elimination of interferences Sample

BCR-723a CW 7b Road dust, 2000 Tunnel dust, 2003 Road dust, 2003

Content of Pd ± S.D. (␮g kg−1 ) (n = 5)

Content of Pt ± S.D. (␮g kg−1 ) (n = 5)

Cation-exchange resin

Cation-exchange resin

Anion-exchange resin

Mathematical correction

7.2 ± 1.0

79.7 ± 8.7 63.7 ± 3.5 99.5 ± 19.6 90.4 (n = 2) 265.6 ± 38.3 (n = 3)

80.3 ± 7.3

81.9 58.2 96.7 89.0 273.3

94.2 (n = 2)

± ± ± ± ±

2.9 5.2 6.7 3.4 23.7 (n = 3)

The content of Pt obtained after mathematical correction method is given for comparison. n: number of independent samples leached with aqua regia; for each one three independent separation processes were performed. a Certified value: 81.3 ± 3.3 ␮g kg−1 Pt; 6.0 ± 1.8 ␮g kg−1 Pd. b Value obtained from interlaboratory comparison: 55.0 ± 8.0 ␮g kg−1 Pt; 4.0 ± 1.2 ␮g kg−1 Pd [37]. c Value determined by TXRF: 32.8 ± 3.8 ␮g kg−1 Pd, n = 3 [12].

B.A. Le´sniewska et al. / Analytica Chimica Acta 564 (2006) 236–242

241

Table 3 Recovery of Pt and Pd determined by ICP-MS from standard solution, BCR-723 and road dust after separation of analytes on Cellex-T (0.2 g) and their elution with 4 ml of 0.1 mol l−1 thiourea in 1 mol L−1 HCl Sample

Model solution (1.6 ng Pt; 1.6 ng Pd) BCR-723 (I) BCR-723 (II) BCR-723 + (16 ng Pt; 8 ng Pd) (I) Road dust, 2000 (I) Road dust, 2000 (II)

pH 4 (n = 3)

pH 1 (n = 2)

Recovery of Pt ± S.D. (%)

Recovery of Pd (%)

Recovery of Pt (%)

Recovery of Pd (%)

84.5 ± 5.5 86.5 ± 5.8 100.3 ± 3.6

>200 >150 >150

115.0 109.1 99.5 87.6

>500 >700 >700 >800

81.3 ± 4.4 99.1 ± 10.3

>500 >500

The samples at pH 4 were initially passed through Dowex 50WX-8. n: number of independent separation processes. I, II: independent samples (0.1 g) leached with aqua regia.

spiked with known concentrations of Pt and Pd were used in initial experiments with GFAAS detection. It must be stressed that the high concentration of matrix ions in digested dust samples can overload the capacity of the Cellex-T column. For this reason and to avoid hydrolysis and precipitation of the matrix ions at pH > 4, causing losses of analytes, a sample clean-up procedure, as described above by means of Dowex 50WX-8 resin, was proposed. The solutions of a sample spiked with Pt (35 ␮g L−1 ) and Pd (12 ␮g L−1 ) was passed through the cationexchange resin and after adjustment of pH to about 4 was loaded on Cellex-T column. The efficiency of Pt and Pd retention on different masses of Cellex-T (0.05–0.25 g) was also studied. The obtained results showed that at least 0.2 g of resin must be used for quantitative retention of analytes from such samples. As HCl is a suitable solvent for ICP-MS measurements it was used in the first instant for the elution of studied metals. However, the recovery (n = 3) of Pt and Pd with 4 mL of 2 mol L−1 HCl was only 36.7 ± 4.2% and 73.5 ± 4.9%, respectively. Then it was tested which amount of thiourea added to the HCl solution would assure effective elution of Pt and Pd. It was found that sufficiently good elution of Pt (90.1 ± 7.6%, n = 3) and Pd (102.9 ± 4.7%, n = 3) was achieved by using 4 mL of 0.1 mol L−1 thiourea solution in 1 mol L−1 HCl. The modified procedure was used for the separation of Pt and Pd from model solution (0.8 ␮g L−1 Pt and Pd) as well as for digested dust or BCR-723 samples. The concentration of both analytes was determined in obtained eluates after fivefold dilution by ICP-MS. Good recovery was achieved in all cases for Pt (Table 3), but in the case of Pd the recovery was exceeded to above >200%, indicating interferences during ICPMS measurements. As the content of yttrium and strontium in all solutions was very low, this cannot be consider as a source of those interferences. In that case, even after careful examination of all possible sources of contamination it was not possible to come to the clear explanation of this phenomenon. An effort to simplify the procedure by omitting the cleanup step was also undertaken. It was observed that the retention of Pt and Pd from solutions of pH 1 was about 90%, when Cellex-T resin was conditioned 0.25 mol L−1 HCl instead of with MQ water [42]. In this case good separation of Pt from Hf was obtained (Table 3). The simplified procedure can be an alternative for elimination of spectral interferences arising from HfO+ species. The recovery of Pd was even higher as in previ-

ous experiments performed by former procedure. Therefore, the correct determination of Pd by ICP-MS in solutions containing thiourea, which passed the column filled with Cellex-T is not possible due to interferences of unidentified origin. Perhaps it is the effect of interaction between the eluent and the resin. The proposed procedure of Pt separation on anion-exchange sorbent Cellex-T allows for effective elimination of matrix interferences from digested road dust samples (pH 1). Its main advantages are effortless sample pre-treatment procedure (no need of pH adjustment), and the total recovery of Pt from the resin with diluted and non-toxic reagent. In this way the application of hot concentrated acids for elution [22–26] or expensive ID technique for compensation of non-quantitative separation [24,25] is avoided. The correctness of the above procedure was confirmed also for the separation and determination of Pd in environmental samples by GFAAS. The limiting factor of its application for Pd determination by ICP-MS was the occurrence of unidentified interferences in column eluates. 4. Conclusions For accurate determination of Pt and Pd in environmental samples effective methods of elimination of spectral interferences are necessary. For simultaneous determination of both, Pt and Pd, chemical separation on ion-exchange resins could be used. Such procedures were optimized according to the ICP-MS requirements. It was found that cation-exchange resin Dowex 50 WX-8 allows effective separation of both analytes from the matrix in a case when samples were leached with aqua regia. The separation of matrix on that cation-exchanger for samples after their total digestion with aqua regia and hydrofluoric acid resulted in incorrect (too high) content of analytes determined by ICP-MS. The retention of analytes on anion-exchanger Cellex-T and their elution with small volume of acidic thiourea solution gave accurate results for Pt determination, while incorrect results were obtained for Pd due to unidentified interferences of column origin. However, such procedure could be used in the case of final determination by GFAAS. References [1] K. Ravindra, L. Bencs, R. van Grieken, Sci. Tot. Environ. 318 (2004) 1.

242

B.A. Le´sniewska et al. / Analytica Chimica Acta 564 (2006) 236–242

[2] B. Gomez, A. Palacios, M. Gomez, J.L. Sanchez, G.M. Morrison, S. Rauch, C. McLeod, R. Ma, S. Caroli, A. Alimonti, F. Petrucci, B. Bocca, P. Schramel, M. Zischka, C. Petterson, U. Wass, Sci. Tot. Environ. 299 (2002) 1. [3] R. Merget, G. Rosner, Sci. Tot. Environ. 270 (2001) 165. [4] M. Moldovan, M.A. Palacios, M.M. Gomez, G. Morrison, S. Rauch, C. McLeod, R. Ma, S. Caroli, A. Alimonti, F. Petrucci, B. Bocca, P. Schramel, M. Zischka, C. Petterson, U. Wass, M. Luna, J.C. Saenz, J. Santamaria, Sci. Tot. Environ. 296 (2002) 199. ˙ [5] B. Godlewska-Zyłkiewicz, B. Le´sniewska, J. Michałowski, A. Hulanicki, Int. J. Environ. Anal. Chem. 75 (1999) 71. ˙ [6] B. Godlewska-Zyłkiewicz, B. Le´sniewska, U. G˛asiewska, A. Hulanicki, Anal. Lett. 33 (2000) 2805. [7] A. Limbeck, J. Rendl, H. Puxbaum, J. Anal. At. Spectrom. 18 (2003) 161. [8] K. Boch, M. Schuster, G. Risse, M. Schwarzer, Anal. Chim. Acta 459 (2002) 257. [9] E. Heinrich, G. Schmidt, K.L. Kratz, Fresenius J. Anal. Chem. 354 (1996) 883. [10] E. Helmers, N. Mergel, Fresenius J. Anal. Chem. 362 (1998) 522. [11] P. Kovacheva, R. Djingova, Anal. Chim. Acta 464 (2002) 7. ˙ [12] B.A. Le´sniewska, B. Godlewska-Zyłkiewicz, B. Bocca, S. Caimi, S. Caroli, A. Hulanicki, Sci. Tot. Environ. 321 (2004) 93. [13] M. Moldovan, M.M. Gomez, M.A. Palacios, J. Anal. At. Spectrom. 14 (1999) 1163. [14] M.B. Gomez, M.M. Gomez, M.A. Palacios, Anal. Chim. Acta 404 (2000) 285. [15] F. Petrucci, B. Bocca, A. Alimonti, S. Caroli, J. Anal. At. Spectrom. 15 (2000) 525. [16] R. Djingova, H. Heidenreich, P. Kovacheva, B. Markert, Anal. Chim. Acta 489 (2003) 245. [17] R. Vlasankowa, V. Otruba, J. Bendl, M. Fisera, V. Kanicky, Talanta 48 (1999) 839. [18] H. Mukai, Y. Ambe, M. Morita, J. Anal. At. Spectrom. 5 (1990) 75. [19] J.D. Whiteley, F. Murray, Sci. Tot. Environ. 317 (2003) 121.

[20] T. Meisel, N. Fellner, J. Moser, J. Anal. At. Spectrom. 18 (2003) 720. [21] E. Higney, V. Olive, A.B. MacKenzie, I.D. Pulford, Appl. Geochem. 17 (2002) 123. [22] S. Hann, G. K¨ollensperger, K. Kanitsar, G. Stingeder, J. Anal. At. Spectrom. 16 (2001) 1057. [23] I. Jarvis, M.M. Totland, K.E. Jarvis, Analyst 122 (1997) 19. [24] M. M¨uller, K.G. Heumann, Fresenius J. Anal. Chem. 368 (2000) 109. [25] K. Kanitsar, G. K¨ollensperger, S. Hann, A. Limbeck, H. Puxbaum, G. Stingeder, J. Anal. At. Spectrom. 18 (2003) 239. [26] K. Akatsuka, J.W. McLaren, in: F. Zereini, F. Alt (Eds.), Anthropogenic Platinum-Group Element Emissions, Their Impact on Man and Environment, Springer-Verlag, Berlin, 2000, p. 123. [27] I. Matsubara, Y. Takeda, K. Ishida, Fresenius J. Anal. Chem. 366 (2000) 213. [28] A.G. Coaedo, M.T. Dorado, I. Padilla, F. Alguacil, Anal. Chim. Acta 340 (1997) 31. [29] F.W.E. Strelow, A. Victor, Anal. Chim. Acta 248 (1991) 535. [30] M. Brzezicka, I. Baranowska, Spectrochim. Acta Part B 56 (2001) 2513. [31] K. Pyrzy´nska, M. Trojanowicz, Crit. Rev. Anal. Chem. 29 (1999) 313. [32] D. Nachtigall, S. Artelt, G. W¨unsch, J. Chromatogr. A 775 (1997) 197. [33] Y.H. Kim, Y. Nakano, Water Res. 39 (2005) 1324. [34] M. Iglesias, E. Antico, V. Salvado, Anal. Chim. Acta 381 (1999) 61. [35] X. Su, M. Wang, Y. Zhang, J. Zhang, H. Zhang, Q. Jin, J. Anal. At. Spectrom. 16 (2001) 1341. ˙ [36] B.A. Le´sniewska, B. Godlewska-Zyłkiewicz, A. Hulanicki, Chem. Anal. (Warsaw) 50 (2005) 945. [37] P. Schramel, M. Zischka, H. Muntau, B. Stojanik, R. Dams, M. Gomez, Ph. Quevauviller, J. Environ. Monit. 5 (2000) 443. [38] M.M. Moldovan, M.M. Gomez, M.A. Palacios, Anal. Chim. Acta 478 (2003) 209. [39] K. Brajter, K. Słonawska, Talanta 30 (1983) 471. [40] K. Brajter, K. Słonawska, Mikrochim. Acta [Wien] 1 (1989) 137. [41] A. Lasztity, A. Kelko-Levai, K. Zih-Perenyi, I. Varga, Talanta 59 (2003) 393. [42] B.A. Le´sniewska, PhD Thesis, University of Białystok, Białystok, 2004.