A Comparative Study of Metal Readsorption in the Application of a Three-Stage Sequential Extraction Scheme and Two Accelerated Versions (Ultrasonic and Single Extractions)

Accelerated methods for partitioning of Cd, Cr, Cu, Ni, Pb and Zn, such as the use of single and ultrasonic extractions are assessed in terms of readsorption and compared with the three-stage sequential extraction scheme (SES) of the Standards Measurements and Testing Programme (SM&T). The standard addition approach was employed for characterising the readsorption artefact by applying the above fractionation methods over different certified reference materials (CRMs), BCR 701, BCR 601 (lake sediments), and BCR 141R soil. Ultrasonic extractions provided higher readsorption, mainly for BCR 141R and 601 as compared to conventional SES, the role of ultrasound in the activation of adsorptive sites being significant. The single extraction approach seemed to be inadequate with samples containing large amounts of carbonates such as BCR 141 R but worked well with both lake sediments. The readsorption phenomena are mainly occurring in the most labile fraction (i.e. acid soluble) and has been observed to be matrix dependent. The extent of such phenomenon is also dependent on the extraction methodology, i.e conventional vs accelerated. INTRODUCTION Since Tessier’s contribution [1] and the Community Bureau of Reference (BCR) harmonization [2], sequential extraction schemes (SESs) have been widely applied and, thus, received an increasing attention during last years to estimate metal mobility and availability in the environment [3]. Several methods for speeding up metal fractionation in order to overcome the time-consuming drawback of SESs, such as application of single extractions or microwave and ultrasonic treatments, have been developed. Single extractions [4] as a simple modification of SESs, i.e. just performing simultaneously the extractions over different aliquots of the same sample, allow diminishing the whole treatment up to a day. The fraction contents, except the first one, are evaluated by subtracting the results obtained in two consecutive stages. On the other hand, ultrasonic versions of SESs have demonstrated their potential for drastic acceleration of conventional SESs, providing similar extractability for several metals in a variety of environmental samples [5-9]. Microwave treatments may be less suitable for shortening the operation time in SESs since the heating could cause significant changes on metal extraction and fractionation, being the labile fractions the most affected [10-13]. The readsorption artefact has been identified as one of the most significant problems affecting the results obtained *Address correspondence to this author at the Universidade de Vigo, Departamento de Química Analítica y Alimentaria (Area de Química Analítica), Facultad de Ciencias (Química), As Lagoas-Marcosende s/n 36200, Vigo, Spain; Tel: +34-986-812281; Fax: +34-986-812556; E-mail: [email protected] by sequential extractions [14] As a result of readsorption, the concentration of trace metals associated with the dissolving phase is underestimated whereas the concentration of trace metals associated with the receiving phase is overestimated. Three different approaches have been employed to assess the extent of readsorption, i.e. use of model synthetic phases, natural sediments spiked with model synthetic phases and the standard addition method. Kheboian and Bauer [15], using model synthetic phases, observed a significant readsorption of Cu, Pb and Zn when the Tessier scheme was applied. Ajayi and Vanloon [16] employed the natural sediments spiked with model synthetic phases with the aim of supplying fresh unoccupied sites for retention of any metals released during extraction. The main readsorption was attributed to Cr, Pb and Zn over iron and manganese oxides and organic matter. However, using a slightly modified second approach over the BCR-SES, Whalley et al. [17] observed that certain model phases e.g. humic acids and ferrihydrite generally released trace metals earlier than expected. Thus, both approaches, employing artificial model phases, differ markedly from natural sediments, thus biasing the system towards trace metal redistribution. The observed drawbacks in the readsorption assessment by the commented approaches are overcome by the standard addition method. This would be an appropriate approach provided that the spiked metal concentrations were below 100% of the natural existing concentrations in the studied samples; otherwise the readsorption problem would be confused with a probably biasing of the system towards metal redistribution [18]. Anyway, the standard addition should not exceed the native concentration in the sample so that the Metal Readsorption in the Application of a Three-Stage Sequential Extraction Scheme The Open Analytical Chemistry Journal, 2008, Volume 2 41 existing equilibrium is not significantly perturbed by the alteration of the extraction conditions. To our knowledge, accelerated fractionation methods applied to obtain fast information about extractable metal contents have not been characterised in respect to readsorption in the SM&T-SES (Standard Measurements and Testing Program, formerly BCR-SES). Therefore, the aim of this work is to study the extent of the readsorption artefact when using, on one hand, ultrasonic extractions, and on the other hand, single extraction methodology to a conclusive comparison with the conventional SM&T-SES. EXPERIMENTAL Apparatus and Reagents Cd, Cr, Cu, Ni, Pb and Zn were determined in both, extracts and aqua regia digests, using a Perkin Elmer inductively coupled plasma optical emission spectrometer (ICPOES-axial view) (model Optima 4300 DV, Norwalk, CT, USA). Matrix-matched standards with extractants were used for calibration. Calibration standards and reagent blanks were frequently reanalysed as samples for quality control purposes. Analytical grade reagents (Panreac and Merck) were used throughout. All extractant, standards, and rinse solutions were made from ultrapure water with a resistivity greater than 18 M ·cm, using a Milli-Q water purification system (Millipore, Molsheim, France). All glassware and plastic containers were previously soaked overnight in 25% nitric acid and rinsed. Sample digestions for pseudototal determination were performed in perfluoroalcoxy (PFA) vessels, with a CEM Corporation microwave laboratory unit (CEM Mars X, Mathews, NC, USA). Ultrasonic agitation was performed using a 100 W, 20 KHz probe sonicator (Sonics and Materials; model VC 100), equipped with titanium probes. Extracts were separated from solid residues using an Alresa centrifuge (model C-5, Barcelona, Spain). A Crison pH-meter (model 2001 micropH) was used for pH adjustments of the extracts. Certified Reference Materials Three certified reference materials (CRMs) were employed in this study for assessing readsorption, a description of their main characteristics being shown in Table 1. Among the considered CRMs, BCR 701 and 601 are both lake sediments, while BCR 141R is a heavy metal contaminated soil. BCR 701, 601 and 141R were obtained from the SM&T of the European Commission. BCRs 701 and 601 are provided with certified/indicative fraction-specific metal concentrations according to the modification of the BCR-SES that lead to the SM&T-SES [19]. No sequential extraction data is available for BCR 141R, which provides only aqua regia soluble contents. Major compositions and total extractable metal contents of the CRMs studied are shown in Table 1. Analytical Procedures Extraction conditions corresponding to each stage of the employed SM&T-SES are shown in Table 2. An ultrasonic version of this SES was applied following previously optimised conditions for treatments with an ultrasonic probe [6]. To properly compare individual results of each fraction, ultrasound extractions were always applied over the remaining residues after removal of the previous fractions using the conventional SES. The single extraction approach followed the same operational conditions than the conventional procedure shown on the left of the table, but over fresh samples in every step. Accelerated and conventional extractions were performed by triplicate, including samples, vessel, reagent and procedural blanks, which were run simultaneously. As a quality control, a mass balance was established by comparing the pseudototal metal content (i.e. applying a microwave digestion with HCl+HNO3) with the sum of the extracted metal percentages in the four steps. After each step of SES, the suspension was centrifuged and the supernatant from the solid phase separated. Extracts were filtered through 0.22micron filter Millex-GS (Millipore, Ireland) in order to avoid the clogging of the nebulizer when using ICP-OES. The resulting extracts were stored in polypropylene bottles refrigerated at 4oC prior to analysis, with the exception of the extracts corresponding to the second fraction, which were analysed immediately due to the instability and degradation of the extractant reagent. In order to compare the relative extractability of the fractionation approaches tried, the term ‘recovery’ was employed as in earlier studies [6,13], which is defined as follows: Table 1. Description of the CRMs Examined in this Study. Major Composition and Total Trace Metals of the CRMs Studied Major Components (%) Total metal (mg·kg -1 ) CRM SiO2 Al2O3 MgO TiO2 Fe2O3 P2O5 K2O Organic C (wt.%) Cd Cr Cu Ni Pb Zn BCR 141R [20] Calcareous loam soil from upper 10cm of fields near Pellegrino, Italy. Ground to < 90 μm. 51.1 11.6 2.3 0.6 4.0 0.4 1.6 11 14.6 195 46.4 103 57.2 283 BCR 701 [21] Lake sediment from different sampling sites of Lake Orta, Italy known for serious metal contamination. Ground to < 90 μm 59.4 15.6 3.2 0.7 6.4 0.5 2.6 10 11.7 272 275 103 143 454 BCR 601 [22] Lake sediment from different sampling sites of Lake Flumendosa, Italy. Collected in March 1994. Ground to < 90 μm. 48.8 13.9 2.2 0.8 7.3 0.9 2.6 5 11 148 240 72 231 824 42 The Open Analytical Chemistry Journal, 2008, Volume 2 Pérez et al. Recovery (%) = [Metal released by ultrasonic or single extractions] / [Metal released by the conventional three-stage SES] 100 Standard Addition Experiments Experiments were addressed to estimate the extent of the readsorption phenomena of each metal (Cd, Cr, Cu, Ni, Pb and Zn) in every fraction of the SM&T-SES. Firstly, the conventional procedure was carried out in order to determine the appropriate metal amounts to be spiked in a particular fraction. Then, a new SES was performed on a separate sample portion after appropriate addition of small volumes (50300 μl) of acidified metal stock solutions. Thus, for F1 readsorption estimation, the spike was added to a fresh sample portion. For F2 and F3, the corresponding spikes were added to the remaining portions generated after F1 and F2 of the conventional SES, respectively. The spikes contained no more than 100% of the “native” extracted amount by the conventional SES, in order to maintain the normal extraction conditions. In order to check for possible bias that could affect the assessment of readsorption, other different source of metal losses than that corresponding to the readsorption artefact were investigated. The blank experiments using the same volume of spiked extractant reagent showed that no losses by adsorption onto the wall of the treatment vessels were present during the sequential extraction experiments. On the other hand, non-significant metal amounts were detected in the rinsing water. Readsorption values (%) were obtained considering the difference between the measured concentrations (μg/mL) of the metal in the extracts corresponding to treatments with non-spiked [M]A and spiked [M]C extractant. This difference was compared with [M]B, which is the concentration ( g/mL) of the element added to the extractant before the treatment. The readsorption (%) is calculated as follows: Readsorption (%) = {1([M]C [M]A) / [M]B} 100 RESULTS AND DISCUSSION Fractionation Patterns Ultrasonic extractions, single extractions and conventional SES are compared. Firstly, CRM BCR 601 and CRM BCR 701 were employed for validation of SM&T-SES. In all cases, a good agreement between certified/indicative and found extractable contents is observed (Table 3). It would be expected that in terms of recovery, single extractions did not significantly change the fractionation patterns of the considered metals and gave a similar performance, since the first fraction is common to the conventional procedure and operational conditions (e.g. extractant concentration, agitation time, temperature, etc) remain unchanged. However, while no significant differences between single and conventional extraction for all the metals in the sediment CRMs are observed when the F2 content is estimated, single extractions provide an efficient extraction only for Cd when applied to the soil CRM. The high lability of this element and its main association with manganese oxides would explain Cd behaviour. For the rest of elements, matrix characteristics would account for the lack of extractability in the single extraction. The high carbonate content of BCR 141R, would probably consume an important part of the extractant reagent, being not enough the H + available for completely dissolve the different components that typically are related to reducible fraction (manganese oxides, crystalline and amorphous iron oxyhidroxydes). With a few exceptions (e.g. Pb in BCR 601), single extraction yielded good recoveries in the oxidable fraction. When total extractable contents are compared (sum of F1+F2+F3) (Table 4), single extractions yielded recoveries in the range 90-110% for all metals in both sediments. However, as expected, low recoveries for total extractable content (<60%) are achieved with the soil sample except for Cd. Table 2. Analytical Conditions and Chemical Reagents for the Conventional SES and Ultrasonic Extractions, Both with the aqua regia Digestion Add-On Step. The Single Extraction Approach Follows the Same Operational Conditions than the Conventional SES But Over Fresh Samples in Every Step. Moisture Content of CRMs was Determined by Drying it at 105oC Until Constant Weight. Chemical Reagents and Conditions Step Fraction Conventional Extraction SM&T-SES Ultrasonic Extraction 1 Acid soluble (F1) 0.5 g of CRM portion, 20 ml 0.11 M acetic acid, shake 16 h. Centrifugation and separate extract from residue at 3000 g for 20 min. 0.5 g of CRM portion, 20 ml 0.11 M acetic acid, 12 min ultrasonic sonication at 50% amplitude. Centrifugation and separate extract from residue at 3000 g for 20 min. 2 Reducible (F2) Add 20 ml 0.5 M NH2OH·HCl (pH 1.5) to step 1 residue, shake 16 h. Centrifugation as in step 1. Add 20 ml 0.5 M NH2OH·HCl (pH 1.5) to step 1 residue, ultrasonic sonication during 9 min at 50% amplitude. Centrifugation as in step 1. 3 Oxidisable (F3) Add 5 ml H2O2 pH (2-3) to sep 2 residue digesting at room temperature during 1 h, heat to 85 ± 2oC for 1 h, add further 5 ml H2O2 and heat to 85 ± 2oC for 1 h; add 25 ml 1 M NH4OAc (pH 2) and shake 16 h. Centrifugation as in step 1. Add 5 ml H2O2 pH (2-3) to sep 2 residue digesting at room temperature during 1 h. Digest with ultrasound during 9 min at 50% amplitude, heat near dryness and add 25 ml 1 M NH4OAc (pH 2), ultrasonic sonication 6 min at 50% amplitude. Centrifugation as in step 1. 4 Residual (F4) Validated microwave-assisted digestion of step 3 residue with 6 ml HClconc, 2 ml HNO3 conc, 1 ml H2O. Also used for pseudototal digestion of 0.25 g of original CRM using the same acid mixture. Metal Readsorption in the Application of a Three-Stage Sequential Extraction Scheme The Open Analytical Chemistry Journal, 2008, Volume 2 43 Ultrasonic treatment caused higher metal amounts of Cr, Cu and Zn (BCR 141 R) and Cr and Pb (BCR 601) to be released in F1. Such release enhancement by ultrasonic cavitation is also observed for Fe (137 mg·kg -1 against 45 mg·kg -1 in conventional treatment for BCR 141R) and Mn as a consequence of the related oxides disaggregation and colloid formation (smaller than 0.22 m) Thus, the observed increase of metal release can be attributed to the enhancement of the contact interface between metal oxides and extractant. A comparison of F1 in the three CRMs assayed reveals that BCR 701 provided the best results as all metals are efficiently recovered, while for the other two CRMs only two metals are similarly extracted (Cd and Ni in BCR 141R; Cd and Cu in BCR 601). In F2, a general underestimation is observed for Ni due to a clear lack of extraction efficiency by the ultrasonic agitation, also observed in the sediment sample for Cr, another typical residual element. The rest of elements present recoveries around 100%. This is an apparent contradiction because of the relatively different concentrations of the metals released on F1 and F2. In this sense, Table 4. Summary of Total Extractable Content (F1+F2+F3) (mg·kg -1 ) Obtained for BCR 141R, BCR 701 and BCR 601 by Conventional SES, Single Extraction and Ultrasonic Extraction CRM/Procedure Cd Cr Cu Ni Pb Zn Conventional 10.9 ± 1.2 49 ± 2 16.1 ± 1.3 59 ± 5 37 ± 2 144 ± 17 Single Extraction 11.9 ± 0.5 30.9 ± 0.8 9.5 ± 0.7 24 ± 3 8 ± 3 87 ± 5 Recovery (%) 110 63 59 41 21 60 Ultrasound 9.4 ± 0.3 46.9 ± 1.0 25.2 ± 1.7 17.0 ± 1.0 40.4 ± 0.8 146 ± 5 B C R 1 4 1 R Recovery (%) 86 95 157 29 109 101 Conventional 10.4 ± 1.5 191 ± 8 220 ± 12 59.9 ± 1.3 132 ± 5 360 ± 10 Single Extraction 10.8 ± 0.5 187 ± 10 213 ± 8 56.1 ± 0.9 138 ± 5 363 ± 14 Recovery (%) 105 96 97 94 105 101 Ultrasound 11.3 ± 0.8 159 ± 5 209 ± 7 48.5 ± 1.5 126 ± 4 322 ± 2 B C R 7 0 1 Recovery (%) 109 83 95 81 95 90 Conventional 9.4 ± 0.3 23.8 ± 0.5 152 ± 3 26.4 ± 0.6 229 ± 16 600 ± 14 Single Extraction 9.0 ± 0.9 27.8 ± 1.4 162 ± 6 29 ± 2 218 ± 6 614 ± 10 Recovery (%) 96 108 106 112 95 102 Ultrasound 9.3 ± 0.7 19.1 ± 0.6 160 ± 7 15.0 ± 0.3 199.7 ± 1.3 445 ± 4 B C R 6 0 1 Recovery (%) 99 80 105 57 87 74 Recovery is calculated using the ratio: [total metal content (F1+F2+F3) extracted using the accelerated (single extractions or ultrasonic agitation) procedure/total metal content extracted by the conventional sequential procedure] 100. Table 3. Indicative and Experimental Extractable Concentrations for Cd, Cr, Cu, Ni, Pb and Zn in BCR 601 and 701 CRMs (mg/kg) Using the Conventional Extraction Procedure