Talanta 80 (2009) 434439

Contents lists available at ScienceDirect

Talanta
journal homepage: www.elsevier.com/locate/talanta

Ultrasound energy focused in a glass probe: An approach to the simultaneous and fast extraction of trace elements from sediments
Silvia Fdez-Ortiz de Vallejuelo, Ana Barrena, Gorka Arana, Alberto de Diego , Juan Manuel Madariaga
Department of Analytical Chemistry, University of the Basque Country, P.O. Box 644, E-48080 Bilbao, Basque Country, Spain

a r t i c l e

i n f o

a b s t r a c t
The 3051 USEPA method (or alternatively, the 3051A) can be considered nowadays as a reference method to extract metals from sediments. However, after microwave heating, the sample must be allowed to cool down, which results in a considerable lengthening of the whole analytical process. Microwave ovens and their maintenance are, in addition, expensive, and its use is relatively dangerous. The use of ultrasound focused energy to assist the extraction of chemicals from solid samples is a safe and relatively cheap technique. In this work we propose a new method to extract simultaneously several elements from sediments using ultrasound energy focused in a glass probe to accelerate the process, and check its possibilities to become an alternative to the EPA3051(A) approach. The optimised procedure allows extracting 13 elements in only 6 min, with similar recoveries and, in general, better repetitivities than the EPA3051. In addition, the suspension is only slightly heated during the leaching process. 2009 Elsevier B.V. All rights reserved.

Article history: Received 24 March 2009 Received in revised form 29 June 2009 Accepted 1 July 2009 Available online 9 July 2009 Keywords: Sediments Trace elements Ultrasound focused energy Sample treatment

1. Introduction Metals, specically arsenic, and their compounds are included in the indicative list of main pollutants of the EU water framework directive [1]. They are present in natural aquatic systems at trace and ultratrace levels [2]. There is a continuous exchange of trace elements between sediments and water in oceans, estuaries, rivers and lakes [3]. This exchange plays an important role in the biogeochemical cycle of these pollutants and usually determines their availability to living organisms [4,5]. In addition, sediments serve as reservoir of trace elements and, due to the long residence times involved, they can be used to investigate historical pollution in selected areas [4,6]. In this context, the need of measuring the concentration of trace metals in sediments becomes evident. The analysis of trace and ultratrace elements in sediments is a long process which involves several steps. Nowadays, extraction of the analytes from the matrix constitutes the bottleneck of the analysis in terms of time, reproducibility and accuracy [7]. A quantitative extraction requires the total dissolution of the sample, which is only attained using hydrouoric acid or any of its derivatives (usually combined with other strong acids and/or oxidants) as extractant. In many cases, however, the determination of total concentrations is not mandatory, and the result obtained after the extraction of the analytes using strong mineral acids such as nitric and hydrochloric acids (or mixtures of them, e.g., aqua regia) is enough for toxicolog-

Corresponding author. Fax: +34 946013500. E-mail address: alberto.dediego@ehu.es (A. de Diego). 0039-9140/$ see front matter 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2009.07.002

ical purposes. The so-called pseudo-total metal content (or acid leachable fraction) is usually well up the bioavailable fraction and has been frequently used, for instance, in environmental monitoring programs [8] and toxicological risk assessment [9]. Solidliquid acid leaching has become, consequently, the most popular technique to extract trace elements from sediments. The conventional hot-plate procedure [10] is long and tedious. The methods recommended by the International Organization for Standardisation (ISO 11466 [11]) and the U.S. Environmental Protection Agency (EPA 3050B [12]) use a reux condenser, mixtures of mineral acids and oxidants, and still take several hours. Different approaches, such as pressure-bombs [13], ultrasound baths [14] and both focused [15] or pressurised [16] microwave ovens, have been investigated to accelerate the process. By far, the last option has been the most popular one over the last 10 years. The EPA method (EPA 3051) describes the microwave assisted HNO3 extraction of sediments for metal analysis [17]. Its alternative, the EPA 3051A [18] allows using HNO3 HCl mixtures to improve the performance for several elements. Microwave assisted leaching, however, is affected by several and serious limitations, such as relatively long extraction time and low reproducibility. The sample is subjected to high temperature and pressure and must be allowed to cool down before going on with the general procedure. Last generation microwave ovens and their maintenance, in addition, are relatively expensive, and must be manipulated by well trained personal. Ultrasound energy has been used to accelerate the extraction of organics, metals and also organometals from environmental solid samples [19,20]. Ultrasound baths provide a simple and cheap way to deliver ultrasound energy. They allow using a large variety of

S. Fdez-Ortiz de Vallejuelo et al. / Talanta 80 (2009) 434439

435

solvents as extractants. However, the irradiation power is poor, the reproducibility is low and the ultrasonic efciency is not homogeneously distributed in the bath [21]. Focused ultrasound systems overcome those problems. The ultrasound energy is focused in the tip of a probe, typically made in titanium, which is immersed in the solution to be irradiated. The irradiation power is up to 100 times higher than that provided by a bath. Only long irradiation times result in a constant but modest increase in temperature of the irradiated system. Use of focused ultrasound produces relatively short, quantitative and reproducible extractions of pollutants from solid samples of different origin [19,22,23]. The use of concentrated strong mineral acids as extractant, however, is not recommended by manufacturers, because the half-life of the titanium probe is considerably reduced. Consequently only a few works, all of them appeared before 2000, have been published on the extraction of metals from sediments using an ultrasonic probe [2426]. In all the cases, short extraction times (less than 5 min) and diluted HNO3 (up to 10%, v/v) were used, and recoveries of 60%, 5469%, 15% and 7595% were obtained, respectively, for Cu, Pb, Cr and Cd. From these results it becomes evident that the impossibility of using concentrated acids seriously limits the applicability of the technique. The commercialisation of a glass probe as an alternative to the classical titanium one is recent. Glass is chemically resistant to strong concentrated acids and becomes an excellent candidate to be used in the extraction of metals from solid samples. There is no restriction regarding irradiation time and acid concentration. In addition, potential contamination of the samples with impurities contained in the titanium probe is eliminated [21]. As a drawback, the ultrasound power and amplitude delivered by a glass probe are limited, due to the low mechanical resistance of glass compared to titanium. There is only one work dealing with the extraction of eight metals from solid samples (sewage sludge) for using industrial ultrasound focused on a glass sonotrode to accelerate the process [27]. To our knowledge, this technique has not been yet used in the analysis of sediments. In this work we check the possibilities of the ultrasound energy focused in a glass probe as an alternative to the microwave assisted leaching of trace elements from sediments. The results obtained have allowed us proposing a very fast procedure (6 min of irradiation time) for the simultaneous extraction of several elements (Al, As, Co, Cr, Cu, Fe, Mg, Mn, Ni, Pb, Sn, V and Zn) from sediments and their simultaneous detection and quantication by ICP/MS.

deionised water and nally left in a 10% nitric acid bath for 24 h. Afterwards, it was thoroughly rinsed with deionised and MilliQ water before use. Chemometric analysis of data was carried out with The Unscrambler 9.2 (Camo Process, Oslo, Norway) and Multisimplex 1.0 (Grabitech, Sundsvall, Sweden). 2.2. Instrumentation Weighing solid samples and preparation of calibrants were carried out by means of an analytical balance AE200 (Mettler Toledo, Columbus, OH, USA) with a precision of 0.00001 g. An HD 2070 Sonopuls Ultrasonic Homogenizer (Bandelin, Germany) equipped with a GM 2070 generator (70 W, 20 kHz), an UW 2070 ultrasonic converter and either a SH 70G horn and MS 73 titanium probe (3 mm) or, alternatively, a SH 70 GQ horn and GS 6 glass probe (6 mm) was used in the extraction step. A variable power setting (0100%) allows controlling amplitude of the delivered ultrasound. A maximum power of 35% is recommended by the manufacturer when using the glass probe. In some experiments, the sample was continuously stirred during extraction by means of an Asincro magnetic stirrer (Selecta, Barcelona, EU). Trace element analysis was carried out with an Elan 9000 ICP/MS (PerkinElmer, Ontario, Canada), provided with Ryton cross-ow nebulizer, Scott-type double pass spray chamber and standard nickel cones. Preparation of calibrants and analysis of samples were done inside a clean room (class 100). Argon (99.999%, Praxair, Spain) was used as carrier gas in the ICP/MS measurements. 2.3. Experimental procedure 2.3.1. Ultrasound assisted extraction of sediments Previously freeze-dried and sieved (<75 m) sediment collected in December 2005 in the estuary of the Nerbioi-Ibaizabal River (Bilbao, Bay of Biscay, Basque Country, N 43 19 04.5 ; W 02 59 35.1 ), NI0512, was used in all the experiments. Design of experiments (DoE) was used to study the inuence of several variables on the extraction efciency and to optimise the extraction process. DoE addresses several advantages when compared to the classical oneexperiment-at-a-time approach: (i) it is possible to investigate simultaneously the effect of several factors as well as interactions among them, (ii) the ratio between obtained information and experimental effort is optimised, and (iii) DoE allows modelling mathematically the system under study. An exact amount of dry sediment (0.20.5 g) was transferred to a polyethylene cylindrical extraction vessel together with an appropriate volume of extractant (1020 mL). In some experiments a stir bar was added to the vessel. The probe (titanium or silicaglass) was placed inside the suspension (0.52 cm), the magnetic stirrer was switched on (when applicable) and sonication was applied at a xed power (10100%) and cycles (19 or continuous) during a xed time (53540 s). To avoid cross-contamination the probe was rinsed several times with MilliQ water and then air dried between sample runs. After sonication, the suspensions were ltered through PVDF 0.45 m lters using plastic syringes. In some experiments, the extract was separated from the residual solid by centrifugation at 2500 rpm during 10 min (2) and the supernatant was ltered through PVDF 0.45 m lters before analysis. Two different dilutions of the extract were prepared in water to measure, respectively, minor and major elements. In both of them HNO3 and HCl concentrations were adjusted to 1% and 0.45% respectively by adding the appropriate amounts of concentrated HNO3 and HCl. All dilutions were made by weight. In both dilutions, Be9 , Sc45 , In115 and Bi209 were added (10 g L1 each) as internal standard before analysis. Blanks were processed in a similar way.

2. Experimental 2.1. Materials and reagents Calibrant solutions of the analytes all together were prepared by weight in 1% HNO3 from individual commercial stock solutions of Al, As, Co, Cr, Cu, Fe, Mg, Mn, Ni, Pb, Sn, V and Zn (1000 mg L1 in 5% HNO3 , Specpure, Alfa Aesar, Ward Hill, MA, USA). Be9 , Sc45 , In115 and Bi209 stock standard solutions (1000 mg L1 in 5% HNO3 , Specpure, Alfa Aesar) were added to blank, standard and sample solutions to yield at 10 g L1 in concentration as internal standard. Preparation of calibrants and analysis of samples were performed in a class 100 clean room. Efciency of the extraction method was checked using the certied reference materials NIST 1646a (estuarine sediment, National Institute of Standards and Technology, Gaithersburg, USA). Hydrochloric acid (37%, Tracepur), acetic acid (100%, Pro analysis) and nitric acid (69%, Tracepur) were purchased from Merck (Darmstadt, Germany). Milli-Q quality water ( < 0.05 S cm1 , Millipore, Billerica, MA, USA) was used throughout. All plastic and crystal materials in contact with calibrants or samples were rst cleaned with water and soap, rinsed with

436 Table 1 Experimental conditions used in the ICP/MS analysis. Nebulization ow Plasma ow Auxiliary ow Sample ow Measured isotopes Radiofrequency power Integration time Replicates

S. Fdez-Ortiz de Vallejuelo et al. / Talanta 80 (2009) 434439

0.94 L min1 15 L min1 1.2 L min1 1 L min1 27 Al, 75 As, 59 Co, 52 Cr, 63 Cu, 57 Fe, 24 Mg, 55 Mn, 60 Ni, 208 Pb, 120 Sn, 51 V, 66 Zn 1000 W 1000 ms 34

2.3.2. Analysis of extracts and calibrants The concentrations of Al, As, Co, Cr, Cu, Fe, Mg, Mn, Ni, Pb, Sn, V and Zn in the dilutions of the extracts and calibrant solutions were simultaneously measured by ICP/MS using the external calibration method with internal standard correction [28]. The experimental conditions are summarised in Table 1. All measurements were run in triplicate. The accuracy of the analysis was checked using the NIST 1640 standard reference material (Natural water, National Institute of Standards and Technology, Gaithersburg, USA) with satisfactory results. 3. Results and discussion 3.1. Inuence of several variables on extraction efciency In a rst approach, the inuence of several variables, e.g., sonication time (t), amplitude of the ultrasound (A), length of probe immersed in the suspension (l), and cycle (cy; cy indicates a pulsed mode; for instance, cy = 4 means a 0.4 s active period followed by a 0.6 s passive period) was studied using different solutions as extractants. In each case, experiments were organised according to a full factorial design with variables at two different levels and four replicates of the central point. Experiments were always run in a random order. About 0.2 g of sediment was weighted in a conical (at) bottom extraction vessel and extraction was performed as described in Section 2, without stirring, under the experimental conditions dened in the design. The suspension was ltered before dilution and ICP/MS analysis. Analysis of variance (ANOVA) of the results was carried out by means of Unscrambler to decide if the variables considered or any combination among them signicantly inuenced the extraction of each element at a 95% condence level. First, acetic acid was checked as extractant using the titanium probe to focus the ultrasound energy. In this case, t (5600 s), A (10100%), l (0.52 cm), cy (19 or continuous) and concentration of acetic acid (cHAc , 0.13 mol dm3 ) were studied. A total of 36 experiments were run as suggested by a full factorial design 25 + 4. ANOVA showed that only the inuence of cHAc was signicant (and positive) for all the elements considered. Irradiation time, t, only affected signicantly in a positive way the extraction of As, Cu and Mn. The rest of variables and combination of variables did not inuence the extraction efciency. In a next step, t (5600 s), A (1035%), cy (19 or continuous) and concentration of nitric acid (cHNO3 , 0.0115 mol dm3 ) were considered (24 + 4, 20 experiments) using HNO3 aqueous solutions as extractant and the silicaglass probe. The probe was immersed 1.5 cm in the suspension. It was concluded that only cHNO3 inuenced positively the extraction of all the elements, being the inuence highly signicant. Like in the case of HAc, irradiation time also affected positively the extraction of several elements, As, Fe, Mg and Mn. The rest of variables and combinations were not signicant. Finally, a 24 + 4 full factorial design with 20 extraction experiments was used to study how t (5600 s), A (1035%), cy (19 or continuous) and proportion of HNO3 in the extractant (%HNO3 ,

0100%) affected the extraction of metals when aqueous mixtures of HNO3 and HCl are used as extractants and the silicaglass probe to focus the ultrasound energy. The probe was immersed 1.5 cm in the suspension, and 69% HNO3 and 36% HCl were used to prepare the extractant solution at the proportion dened in the design, in a way that %HNO3 = 0, 50 and 100 results, respectively, in an aqueous solution of HCl 13.2 mol dm3 , an aqueous mixture of HNO3 7.1 mol dm3 and HCl 6.6 mol dm3 , and an aqueous solution of HNO3 14.2 mol dm3 . ANOVA showed that it is possible to dene two groups of elements: those which are better extracted with HNO3 rich solutions (Al, Co, Cu, Mg, Pb and Zn), and those which prefer HCl rich solutions (As, Fe, Mn, Ni, Sn and V). Extraction of Cr is not clearly inuenced by this variable. In addition, extractions of Al, As, Co, Cu, Fe, Mg, Mn and Ni are positively affected by irradiation time, t. The rest of variables and combination showed no clear trend. For each design, the concentration of each element in the leachate was estimated under the best possible experimental conditions using the mathematical models provided by Unscrambler after ANOVA of the results. Concentrations were compared with those obtained after analysis of the sediment according to the EPA 3051 method [17] with a Mars Xpress microwave oven from CEM Corporation, Matthews, NC, USA. The extraction efciencies calculated in this way are shown in Fig. 1a. The best recoveries were always obtained with a mixture of HNO3 HCl as extractant, except for Fe and Mg. In those cases, extraction with HAc offered better results, probably due to the fact that the use of the titanium probe allowed delivering ultrasound energy of higher amplitude. To check the effect of agitation on extraction efciency, analysis of the NI0512 sediment was repeated with (n = 4) and without (n = 4) magnetic stirring. About 0.4 g of sample was weighed in a at bottom extraction vessel, a magnetic bar was added (in experiments with agitation) and it was processed as described before using 20 mL of HNO3 2.8 mol dm3 HCl 3.1 mol dm3 as extractant and the silicaglass probe. The leachate was separated from the solid residue by ltration. The rest of variables was xed as follows: t = 1800 s, A = 35%, l = 1.5 cm and cy = 5. As it can be observed in Fig. 1b, extraction efciency was systematically higher in experiments with agitation. Similar experiments showed that separating the leachate from the solid residue by ltration or centrifugation did not signicantly affect the result of the analysis. Consequently, in all the next experiments the suspension was magnetically stirred during extraction and the leachate was separated by ltration. 3.2. Optimisation of the extraction step After the preliminary experiments described in the previous section it was concluded that the variables with higher inuence on the extraction of the elements considered from sediments using ultrasound energy focused in a silicaglass probe were irradiation time, t, as well as proportion, %HNO3 , and concentration, cacid , of acids in the HNO3 HCl aqueous mixture used as extractant. A central composite design (CCD) with three variables at ve different levels and four repetitions of the central point (18 experiments run in a random order) was used to nd the optimal experimental conditions for the extraction of each element considered. t, %HNO3 and cacid were varied from 0.5 to 59 min, 892% and 0.0111.8 mol L1 , respectively ( to + levels). In all the experiments, about 0.4 g of the NI0512 sediment was suspended in about 20 mL of extractant, the suspension was magnetically stirred along the extraction process, the glass probe was immersed 1.5 cm in the solution and ultrasound amplitude of 35% was applied with a cycle of 5. The extract was separated from the solid residue by ltration. Regression analysis of the obtained results allowed dening the response surface for each element, and observation of these surfaces led to the best experimental conditions summarised in Table 2.

S. Fdez-Ortiz de Vallejuelo et al. / Talanta 80 (2009) 434439

437

Fig. 1. Extraction efciencies (as percentage referred to the EPA 3051 method) found for each element considered: (a) using the combination HAc/titanium (white), HNO3 /silicaglass (dark grey) and HNO3 HCl/silicaglass (black) as extractant/probe. The sediment used in the experiments was the NI0512 collected in the estuary of the Nerbioi-Ibaizabal River in December 2005; (b) after analysis of the same sediment with (n = 4, black) or without (n = 4, white) stirring the suspension during ultrasound assisted extraction with a silicaglass probe and HNO3 HCl as extractant.

The extraction efciencies calculated at those conditions are shown in the last row of Table 2, using as reference the EPA 3051 method (Mars Xpress unit from CEM was used).

Table 3 Extraction efciency (%, compared with the EPA 3051 method) estimated for the elements considered at different experimental conditions. t: irradiation time; %HNO3 : proportion of HNO3 in the HNO3 HCl used as extractant; cacid : concentration of the HNO3 and HCl aqueous solution used to prepare the solution used as extractant. t = 6 min; %HNO3 = 45%; cacid = 6.8 mol L1 Al As Co Cr Cu Fe Mg Mn Ni Pb Sn V Zn 22 95 56 72 67 45 56 98 68 107 88 48 119 2 9 6 5 3 5 5 10 5 6 5 5 7 t = 0.5 min; %HNO3 = 8%; cacid = 8.2 mol L1 17 100 61 71 56 39 46 83 70 90 98 54 106 2 10 6 5 2 5 4 9 5 5 6 3 6 t = 0.5 min; %HNO3 = 55%; cacid = 6.3 mol L1 21 96 57 72 66 43 54 94 69 105 90 49 117 2 9 6 5 3 5 5 9 5 6 6 2 7

3.3. Best conditions for a simultaneous extraction of the elements considered The experimental conditions that provide the best extraction efciency substantially differs from element to element, as observed in Table 2. One of the goals of this work was to propose a method for the simultaneous analysis of trace elements in sediments so, the next step consisted on looking for a combination of t, %HNO3 , and cacid able to provide an equilibrated, and as high as possible, extraction efciency for all the elements considered. MultiSimplex helped us in this task. MultiSimplex is a modication of the Simplex algorithm [29] that allows optimizing simultaneously more than one response (extraction efciency of each element) by changing the values of multiple variables (t, %HNO3 , and cacid ). In this case, the optimisation process was simulated using the mathematical models provided by The Unscrambler (after regression analysis of the results obtained in the CCD, previous section) to calculate the extraction efciency obtained at each combination of variables proposed by MultiSimplex. The same statistical weight was assigned to all the elements. The best set of variables found was as follows: t = 6 min, %HNO3 = 45% and cacid = 6.8 mol L1 . The extraction efciencies estimated at these conditions are shown in the rst column of Table 3. As observed, they are only slightly lower than those provided in Table 2. For many elements the recoveries are close (As, Mn and Sn) or even over (Pb and Zn) those obtained by the EPA 3051 method. Other ones (Co, Cr, Cu, Mg, and Ni) show acceptable recoveries (55%75%) and only for Al, Fe and V the recovery is below 50%.

An USF based extraction strategy will be a valid alternative to the microwave classical approach only if the method is effective and comparatively fast. The second and third columns of Table 3 show the extraction efciencies estimated using an irradiation time as short as 0.5 min and two different acid mixtures as extractant (HNO3 poor mixture (8%) and HNO3 rich mixture (55%)). In general, the extraction efciencies obtained in each case are very similar to those reported in the rst column of the Table, especially in the case of the HNO3 rich mixture. This result suggests that the inuence of irradiation time on the extraction efciency is important only at the very beginning of the extraction process (rst 0.51 min), reaching a plateau very fast.

Table 2 Best experimental conditions found for the ultrasound assisted extraction of each element considered. t: irradiation time (min); %HNO3 : proportion of HNO3 in the HNO3 HCl used as extractant; cacid : concentration (mol L1 ) of the HNO3 and HCl aqueous solutions used to prepare the extractant. EE: extraction efciency (compared with the EPA 3051 method). Al t %HNO3 cacid EE 31 66 8.4 23 As 2 8 8.3 100 Co 59 8 8.5 61 Cr 58 54 7.0 73 Cu 30 53 6.9 70 Fe 29 44 8.5 46 Mg 30 52 7.1 59 Mn 29 45 7.4 104 Ni 1 10 8.0 70 Pb 30 47 6.6 112 Sn 59 92 6.5 99 V 21 92 6.1 45 Zn 29 47 6.7 122

438

S. Fdez-Ortiz de Vallejuelo et al. / Talanta 80 (2009) 434439

Fig. 2. Comparison between the results obtained after the analysis of the NI0512 sediment and the NIST 1646a certied reference material (estuarine sediment) using the ultrasound assisted method proposed in this work (t = 6 min, %HNO3 = 45%, cacid = 6.8 mol L1 , black) and the EPA 3051 method (white). Error bars represent the standard deviation of 8 replicate analyses.

Table 4 Recoveries (calculated with respect to certied values) obtained after the analysis of the NIST 1646a reference material according to the EPA 3051 method (n = 8). Al Mean recovery (%) Standard deviation 38.2 4.8 As 87.1 3.0 Co 93.8 1.8 Cr 53.7 2.8 Cu 118.5 5.1 Fe 78.1 4.8 Mg 77.4 2.8 Mn 52.5 5.0 Ni 64.3 1.4 Pb 60.9 1.9 Sn 146.8 10.3 V 18.0 0.5 Zn 93.3 10.4

3.4. Validation of the proposed method 3.4.1. Accuracy The extraction efciencies provided in Tables 2 and 3 are result of an estimation using mathematical models. In order to compare experimentally the efciencies of the methods, the NI0512 sediment and certied reference material (NIST 1646a) sediment were analysed using the ultrasound based method (n = 8) proposed here (t = 6 min, %HNO3 = 45%, cacid = 6.8 mol L1 ) and the EPA 3051 method (n = 8). This time the MW extractions were carried out using an Multiwave 3000 microwave oven from Anton Paar. Previous experiments demonstrated that the results obtained with the Anton Paar oven did not differ signicantly (95% condence level) from those obtained with the CEM unit. The results obtained are compared in Fig. 2. The recoveries (related to certied total concentrations) obtained after the analysis of the NIST 1646a with the EPA 3051 are also provided in Table 4. No signicant difference between both methods (MW vs USF) was observed, except in the case of tin. It is worth noting that only a non-certied value is provided for tin in the certicate of the NIST 1646a material. 3.4.2. Uncertainty The uncertainty of the optimised ultrasound based method (t = 6 min, %HNO3 = 45%, cacid = 6.8 mol L1 ) was checked by analysis of the NI0512 sediment in 4 consecutive days. In addition, in the rst day four replicates of the analysis were carried out. For comparability, analysis of the same sediment was repeated eight times the same day by the EPA 3051 method (Anton Paar unit). The results, expressed as relative standard deviations are shown in Table 5. As it can be observed, the precision obtained by the USF method was considerably better in most of the cases, with RSDs always below 5%.

3.4.3. Detection limit The detection limit of the method was estimated after replicate analysis (n = 8) of a blank. The results obtained for each element, expressed as concentration of element in the sediment in mg kg1 (for 0.5 g of sediment processed), are shown in Table 5. It is worth noting that the detection limit is higher when the measurement by ICP/MS is done in the second dilution rather than in the rst one (this is the case when the concentration of the element in the extract is higher than about 100 mg L1 ).

Table 5 Uncertainty (calculated after replicate analysis of the NI0512 sediment) of the USF based proposed method and the MW assisted EPA 3051, together with the detection limits for the USF proposed method (calculated after analysis of 8 blanks). Uncertainty (RSD, %) USF Within day (n = 4) Al As Co Cr Cu Fe Mg Mn Ni Pb Sn V Zn 2.0 1.2 2.2 2.3 0.7 3.3 2.1 3.2 1.5 0.9 1.2 1.8 2.0 Between days (n = 4) 2.9 3.7 4.0 2.5 1.7 4.2 3.3 4.7 2.6 1.1 2.2 3.5 2.0 MW Within day (n = 8) 11.2 5.1 3.0 12.8 3.1 3.2 7.9 3.0 5.4 4.9 13.6 10.1 4.7 18 0.9 0.2 0.5 23 12 6.0 0.9 4.0 1.0 1.9 0.3 4.0 Detection limit (mg kg1 ) USF

S. Fdez-Ortiz de Vallejuelo et al. / Talanta 80 (2009) 434439

439

4. Conclusions Ultrasound energy focused in a glass probe is a good alternative to accelerate the simultaneous extraction of elements from sediments. The process is highly repetitive and extremely fast. The recoveries obtained are comparable to those of the microwave assisted EPA 3051 method. In addition, the suspension of the sediment in the extractant is only slightly heated during the process. As a consequence: (i) possible losses of analytes by evaporation are minimised and (ii) the extract must not be allowed to cool down before going on with the general analysis procedure. The equipment involved and its maintenance is cheap, and operation safe and extremely simple. As a drawback of the technique, it should be mentioned that samples must be treated one by one but, since the analyst can go on with the general procedure as soon as the extraction process is nished, sample throughput is similar to that of a standard microwave oven. The method proposed here is appropriate for monitoring purposes, ecotoxicological studies and risk assessment. Acknowledgment This work has been nancially supported by the ETORTEK Programme of the Basque Government through the BERRILUR II project (Ref. IE06-179). References
[1] European Commission Off., J. Eur. Commun. L327 (2000) 172. [2] C. Reimann, P. de Caritat, Chemical Elements in the Environment, 1st ed., Springer-Verlag, Berlin, 1998. [3] G.A. Van Den Berg, J.P.G. Loch, L.M. Van Der Heijdt, J.J.G. Zwolsman, Trace ElementsTheir Distribution and Effects in the Environment in: B. Marker, K. Friese (Eds.), Trace Metals in the Environment Series, vol. 4, Elsevier, Amsterdam, 2000, pp. 517533.

[4] J.E. Rae, Biogeochemistry of Intertidal Sediments in: T.D. Jickells, J.E. Rae (Eds.), Cambridge Environmental Chemistry Series, vol. 9, Cambridge University Press, Cambridge, 1997, pp. 1641. [5] W.J.G.M. Peijnenburg, M.G. Vijver, Pure Appl. Chem. 79 (2007) 23512366. [6] M.J. Charles, R.A. Hites, Sources and Fates of Aquatic Pollutants, in: R.A. Hites, S.J. Eisenreich (Eds.), Advances in Chemistry Series, vol. 216, Oxford University Press, Oxford, 1987, pp. 365389. [7] J.M. Cook, M.J. Gardner, A.H. Grifths, M.A. Jessep, J.E. Ravenscroft, R. Yates, Mar. Pollut. Bull. 34 (1997) 637644. [8] M. Jayaprakash, M.P. Jonathan, S. Srinivasalu, S. Muthuraj, V. Ram-Mohan, N. Rajeshwara-Rao, Estuar. Coastal Shelf Sci. 76 (2008) 692703. [9] A. Sahuquillo, A. Rigol, G. Rauret, TRAC Trend. Anal. Chem. 22 (2003) 152159. [10] N.R. McQuaker, D.F. Brown, P.D. Kluckner, Anal. Chem. 51 (1979) 10821084. [11] International Organization for Standardisation, Soil quality, extraction of trace elements soluble in aqua regia, ISO 11466, 1995. [12] U.S. Environmental Protection Agency (EPA), Method 3050B, Acid Digestion of Sediments, Sludges and Oils, 1996. [13] J.C. Woo, D.M. Moon, H. Kawaguchi, Anal. Sci. 12 (1996) 195200. [14] H. Gngr, A. Elik, Microchem. J. 86 (2007) 6570. [15] J. Sanz, A. de Diego, J.C. Raposo, J.M. Madariaga, Anal. Chim. Acta 508 (2004) 107117. [16] A. Robbat Jr., R.L. Simpson, Fresenius J. Anal. Chem. 364 (1999) 305312. [17] U.S. Environmental Protection Agency (EPA), Method 3051, Microwave Assisted Acid Digestion of Sediments, Sludges, Soils and Oils, 1994. [18] U.S. Environmental Protection Agency (EPA), Method 3051A, Microwave Assisted Acid Digestion of Sediments, Sludges, Soils and Oils, 2007. [19] J.L. Capelo, A.M. Mota, Curr. Anal. Chem. 1 (2005) 193201. [20] M. Zabaljauregui, A. Delgado, A. Usobiaga, O. Zuloaga, A. de Diego, J.M. Madariaga, J. Chromatogr. A 1148 (2007) 7885. [21] H.M. Santos, J.L. Capelo, Talanta 73 (2007) 795802. [22] M. Gallego-Gallegos, M. Liva, R.M. Olivas, C. Cmara, J. Chromatogr. A 1114 (2006) 8288. [23] J. Sanz-Landaluze, L. Bartolom, O. Zuloaga, L. Gonzlez, C. Dietz, C. Cmara, J. Anal. Atom. Spectrom. 384 (2006) 13311340. [24] J.L. Capelo, I. Lavilla, C. Bendicho, J. Anal. Atom. Spetrom. 13 (1998) 12851290. [25] L. Amoedo, J.L. Capelo, I. Lavilla, C. Bendicho, J. Anal. Atom. Spetrom. 14 (1999) 12211226. [26] E.C. Lima, F. Barbosa Jr., F.J. Drug, M.M. Silva, M.G.R. Vale, J. Anal. Atom. Spetrom. 15 (2000) 9951000. [27] D. Hristozov, C.E. Domini, V. Kmetov, V. Stefanova, D. Georgieva, A. Canals, Anal. Chim. Acta 516 (2004) 187196. [28] S. Fernndez, U. Villanueva, A. de Diego, G. Arana, J.M. Madariaga, J. Mar. Sys. 72 (2008) 332341. [29] J.A. Nelder, R. Mead, Comp. J. 7 (1965) 308313.