Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Spectrochimica Acta Part B 58 (2003) 1177–1233 Review Solid phase extraction of trace elements ´ Camel* Valerie Institut National Agronomique Paris-Grignon, Analytical Chemistry Laboratory, 16 Rue Claude Bernard, Paris Cedex 05 75231, France Received 2 December 2002; accepted 26 March 2003 Abbreviations: AAS, Atomic absorption spectrometry; ACN, Acetonitrile; AES, Atomic emission spectrometry; APDC, Ammonium pyrrolidine dithiocarbamate; BAS, Biswl-hydroxy-9,10-anthraquinone-2-methylxsulfide; 5-BrPADAP 2-(5-bromo-2pyridylazo)-5-diethylaminophenol; BSQ, 8-(Benzenesulfonamido)quinoline; 18C6, 18-Crown-6; CA, Chromotropic acid; CTA, Cetyltrimethylammonium; CV-AAS, Cold vapour atomic absorption spectrometry; DAD, Diode array detector; DDQ, 7-Dodecenyl8-quinolinol; DDTC, Diethyldithiocarbamate; DDTP, 0,0-Diethyl-dithiophosphate; DEBT, N,N9-Diethyl-N9-benzoylthiourea; DESe, Diethyl selenide; DMBS, Dimethylglyoxal bis(4-phenyl-3-thiosemicarbazone); DMDSe, Dimethyl diselenide; DMG, Dimethylglyoxime; DMSe, Dimethyl selenide; DPC, Diphenylcarbazide; DPCO, Diphenylcarbazone; DPD, N,N-Dimethyl-p-phenylenediamine; DPTH, l,5-bis(di-2-pyridyl) methylene dithiocarbohydrazide; DVB-VP, Divinylbenzene-vinylpyrrolidone; DZ, Dithizone; DzS, Dithizone sulfonic acid; ECD, Electron capture detection; EDTA, Ethylene diamine tetraacetic acid; ERT, Eriochrome black-T; ETAAS, Electrothermal atomic absorption spectrometry; Et, Ethyl; EtOH, Ethanol; F-AAS, Flame atomic absorption spectrometry; FI, Flow injection; FPD, Flame photometric detection; FZ, Ferrozine; GC, Gas chromatography; GCB, Graphitized carbon black; HDEHP, Bis(2-ethylhexyl) hydrogen phosphate; HMDC, Hexamethylenedithiocarbamate; H2 MEHP, 2-Ethylhexyl dihydrogen phosphate; 8-HQ, 8-Hydroxyquinoline; 8-HQ-5-SA, 8-Hydroxyquinoline-5-sulfonic acid; HT18C6, Hexathia-18-crown-6; HT18C6TO, Hexathia-18-crown-6-tetraone; IBMK, Isobutyl methyl ketone; ICP, Inductively coupled plasma; IDA, Iminodiacetate; IP, Ion-pair; KR, Knotted reactor; LC, Liquid chromatography; LLE, Liquid–liquid extraction; LOD, Limit of detection; MBT, 2Mercaptobenzothiazole; Me, Methyl; MeOH, Methanol; MPSP, 3-Methyl-l-phenyl-4-stearoyl-5-pyrazolone; MPT, Microwave plasma torch; MS, Mass spectrometry; NCH, Neocuproine; NDSA, 2-Naphthol-3,6-disulfonic acid; NN, 1-Nitroso-2-naphthol; ODETA, 4(N-octyl)diethylenetriamine; PA, Polyacrylate; PAA, Piconilic acid amide; PADMAP 2-(2-pyridylazo)-5-dimethylaminophenol; PAN, 1-(2-pyridylazo)2-naphthol; PaPhA, Poly(aminophosphonic acid); PAR, 4-(2-pyridylazo)resorcinol; PC, Pyrocatechol; PDATA, Propylenediaminetetraacetic acid; PDT, 3-(2-Pyridyl)-5,6-diphenyl-l,2,4-triazine; PDTC, Poly(dithiocarbamate); PE, Polyethylene; PGC, Porous graphitized carbon; Ph, Phenyl; PipDTC, Piperidine dithiocarbamate; PS-DVB, Polystyrene-divinylbenzene; PTFE, Polytetrafluoroethylene; PUF, Polyurethane foam; PV, Pyrocatechol violet; ROMP, Ring-opening metathesis polymerisation; SA, Salicylic acid; SDS, Sodium dodecylsulfate; SFE, Supercritical fluid extraction; SGBM, Silica gel bound macrocycles; SPE, Solid phase extraction; SPS, Solid-phase spectrophotometry; TAN, 1-(2-tiazolylazo)-2-naphthol; TBP, Tri-n-butyl phosphate; T4BPP, Tetra-(4-bromophenyl)-porphyrin; TBT, Tributyltin; TCPP, Carboxyphenylporphyrin; THF, Tetrahydrofuran; TOPO, Tri-noctylphosphine oxide; TPhT, Triphenyltin; TS, Methylthiosalicylate; TSA, Thiosalicylic acid; TTA, 2-Thenoyltrifluoroacetone; UV, Ultraviolet; XO, Xylenol orange. *Corresponding author. Tel.: q33-1-44-08-17-25; fax: q33-1-44-08-16-53. E-mail address: [email protected] (V. Camel). 0584-8547/03/$ - see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0584-8547Ž03.00072-7 1178 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 1. Introduction Despite the selectivity and sensitivity of analytical techniques such as atomic absorption spectrometry, there is a crucial need for the preconcentration of trace elements before their analysis due to their frequent low concentrations in numerous samples (especially water samples). Additionally, since high levels of non-toxic components usually accompany analytes, a clean-up step is often required. Liquid–liquid extraction is a classical method for preconcentrating metal ions andyor matrix removal. Solid phase extraction (SPE) is another approach that offers a number of important benefits. It reduces solvent usage and exposure, disposal costs and extraction time for sample preparation. Consequently, in recent years SPE has been successfully used for the separation and sensitive determination of metal ions, mainly in water samples. After outlining the theory of this technique, guidelines are given for the development of SPE-based methods for preconcentration of many trace elements. Finally, examples of applications are presented. 2. Theory The principle of SPE is similar to that of liquid– liquid extraction (LLE), involving a partitioning of solutes between two phases. However, instead of two immiscible liquid phases, as in LLE, SPE involves partitioning between a liquid (sample matrix) and a solid (sorbent) phase. This sample treatment technique enables the concentration and purification of analytes from solution by sorption on a solid sorbent. The basic approach involves passing the liquid sample through a column, a cartridge, a tube or a disk containing an adsorbent that retains the analytes. After all of the sample has been passed through the sorbent, retained analytes are subsequently recovered upon elution with an appropriate solvent. The first experimental applications of SPE started fifty years ago w1,2x. However, its growing development as an alternative approach to liquid–liquid extraction for sample preparation started only in the mid-1970s. It has been extensively used in the past fifteen years for the preconcentration of organic micropollu- tants, especially pesticides, in water samples w3x. However, numerous studies have also shown the great potential of this technique for speciation studies. 2.1. Presentation of the technique 2.1.1. Basic principles An SPE method always consists of three to four successive steps, as illustrated in Fig. 1. First, the solid sorbent should be conditioned using an appropriate solvent, followed by the same solvent as the sample solvent. This step is crucial, as it enables the wetting of the packing material and the solvation of the functional groups. In addition, it removes possible impurities initially contained in the sorbent or the packaging. Also, this step removes the air present in the column and fills the void volume with solvent. The nature of the conditioning solvent depends on the nature of the solid sorbent. Typically, for reversed phase sorbent (such as octadecyl-bonded { silica), methanol is frequently used, followed with water or aqueous buffer whose pH and ionic strength are similar to that of the sample. Care must be taken not to allow the solid sorbent to dry between the conditioning and the sample treatment steps, otherwise the analytes will not be efficiently retained and poor recoveries will be obtained. If the sorbent dries for more than several minutes, it must be reconditioned. The second step is the percolation of the sample through the solid sorbent. Depending on the system used, volumes can range from 1 ml to 1 l. The sample may be applied to the column by gravity, pumping, aspirated by vacuum or by an automated system. The sample flow-rate through the sorbent should be low enough to enable efficient retention of the analytes, and high enough to avoid excessive duration. During this step, the analytes are concentrated on the sorbent. Even though matrix components may also be retained by the solid sorbent, some of them pass through, thus enabling some purification (matrix separation) of the sample. The third step (which is optional) may be the washing of the solid sorbent with an appropriate solvent, having a low elution strength, to eliminate matrix components that have been retained by the V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 1179 Fig. 1. SPE operation steps. solid sorbent, without displacing the analytes. A drying step may also be advisable, especially for aqueous matrices, to remove traces of water from the solid sorbent. This will eliminate the presence of water in the final extract, which, in some cases, may hinder the subsequent concentration of the extract andyor the analysis. The final step consists in the elution of the analytes of interest by an appropriate solvent, without removing retained matrix components. The solvent volume should be adjusted so that quantitative recovery of the analytes is achieved with subsequent low dilution. In addition, the flow-rate should be correctly adjusted to ensure efficient elution. It is often recommended that the solvent volume be fractionated into two aliquots, and before the elution to let the solvent soak the solid sorbent. 2.1.2. Retention of trace elements on the sorbent Adsorption of trace elements on the solid sorbent is required for preconcentration (see Fig. 2). The mechanism of retention depends on the nature of the sorbent, and may include simple adsorption, chelation or ion-exchange. Also, for trace elements, ion-pair solid phase extraction may be used. 2.1.2.1. Adsorption. Trace elements are usually adsorbed on solid phases through van der Waals forces or hydrophobic interaction. Hydrophobic interaction occurs when the solid sorbent is highly non-polar (reversed phase). The most common sorbent of this type is octadecyl-bonded silica (C18-silica). More recently, reversed polymeric phases have appeared, especially the styrene-divinylbenzene copolymer that provides additional pp interaction when p-electrons are present in the analyte w4x. Elution is usually performed with organic solvents, such as methanol or acetonitrile. Such interactions are usually preferred with online systems, as they are not too strong and thus they can be rapidly disrupted. However, because most trace element species are ionic, they will not be retained by such sorbents. 2.1.2.2. Chelation. Several functional group atoms are capable of chelating trace elements. The atoms most frequently used are nitrogen (e.g. N present in amines, azo groups, amides, nitriles), oxygen (e.g. O present in carboxylic, hydroxyl, phenolic, ether, carbonyl, phosphoryl groups) and sulfur (e.g. S present in thiols, thiocarbamates, thioethers). The nature of the functional group will give an idea of the selectivity of the ligand towards 1180 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 trace elements. In practice, inorganic cations may be divided into 3 groups: – group I-‘hard’ cations: these preferentially react via electrostatic interactions (due to a gain in entropy caused by changes in orientation of hydration water molecules); this group includes alkaline and alkaline-earth metals (Ca2q, Mg2q, Na2q) that form rather weak outer-sphere complexes with only hard oxygen ligands. – group II-‘borderline’ cations: these have an intermediate character; this group contains Fe2q, Co2q Ni2q Cu2q Zn2q Pb2q Mn2q. They possess affinity for both hard and soft ligands. – group III-‘soft’ cations: these tend to form covalent bonds. Hence, Cd2qand Hg2q possess strong affinity for intermediate (N) and soft (S) ligands. Fig. 2. Interactions occurring at the surface of the solid sorbent. F, functional group; TE, trace element; MS, matrix solvent; MI, matrix ions; ES, elution solvent. For soft metals, the following order of donor atom affinity is observed: 0-N-S. A reversed order is observed for hard cations. For a bidentate ligand, affinity for a soft metal increases with the overall softness of the donor atoms: (0, 0)-(0, N)-(N, N)-(N, S). The order is reversed for hard metals. In general, the competition for a given ligand essentially involves Group I and Group II metals for O sites, and metals of Group II and Group III for N and S sites. The competition between metals of Group I and Group III is weak. Chelating agents may be directly added to the sample for chelating trace elements, the chelates being further retained on an appropriate sorbent. An alternative is to introduce the functional chelating group into the sorbent. For that purpose, three different means are available: (1) the synthesis of new sorbents containing such groups (new sorbents); (2) the chemical bonding of such groups on existing sorbents (functionalized sorbents); and (3) the physical binding of the groups on the sorbent by impregnating the solid matrix with a solution containing the chelating ligand (impregnated, coated or loaded sorbents). The latter remains the most simple to be used in practice. Its main drawback is the possible flush of the chelating agent out of the solid sorbent during sample percolation or elution that reduces the lifetime of the impregnated sorbent. V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Different ligands immobilized on a variety of solid matrices have been successfully used for the preconcentration, separation and determination of trace metal ions. Chelating agents with an hydrophobic group are retained on hydrophobic sorbents (such as C18-silica). Similarly, ion-exchange resins are treated with chelating agents containing an ionexchange group, such as a sulfonic acid derivative of dithizone (i.e. diphenylthiocarbazone) (DzS), 5-sulfo-8-quinolinol, 5-sulfosalicylic acid, thiosalicylic acid, chromotropic acid, or carboxyphenylporphyrin (TCPP) w5–8x. Binding of metal ions to the chelate functionality is dependent on several factors: (1) nature, charge and size of the metal ion; (2) nature of the donor atoms present in the ligand; (3) buffering conditions which favor certain metal extraction and binding to active donor or groups; and (4) nature of the solid support (e.g. degree of cross-linkage for a polymer). In some cases, the behavior of immobilized chelating sorbents towards metal preconcentration may be predicted using the known values of the formation constants of the metals with the investigated chelating agent w9x. However, the presence of the solid sorbent may also have an effect and lead to the formation of a complex with a different stoichiometry than the one observed in a homogeneous reaction w10,11x. In fact, several characteristics of the sorbent should be taken into account, namely the number of active groups available in the resin phase w7,10x, the length of the spacer arm between the resin and the bound ligand w12x, and the pore dimensions of the resin w13x. 2.1.2.3. Ion-pairing. When a non-polar sorbent is to be used, an ion-pair reagent (IP) can be added to the sorbent w14x. Such reagents contain a nonpolar portion (such as a long aliphatic hydrocarbonated chain) and a polar portion (such as an acid or a base). Typical ion-pair reagents are quaternary ammonium salts and sodium dodecylsulfate (SDS) w15,16x. The non-polar portion interacts with the reversed-phase non-polar sorbent, while the polar portion forms an ion-pair with the ionic species present in the matrix (that could be either free metallic species in solution or complexes). 1181 2.1.2.4. Ion exchange. Ion-exchange sorbents usually contain cationic or anionic functional groups that can exchange the associated couter-ion. Strong and weak sites refer to the fact that strong sites are always present as ion-exchange sites at any pH, while weak sites are only ion-exchange sites at pH values greater or less than the pKa. Strong sites are sulfonic acid groups (cation-exchange) and quaternary amines (anion-exchange), while weak sites consist of carboxylic acid groups (cation-exchange) or primary, secondary and tertiary amines (anion-exchange). These groups can be chemically bound to silica gel or polymers (usually a styrene-divinylbenzene copolymer), the latter allowing a wider pH range. An ion-exchanger may be characterized by its capacity, resulting from the effective number of functional active groups per unit of mass of the material. The theoretical value depends upon the nature of the material and the form of the resin. However, in the column operation mode, the operational capacity is usually lower than the theoretical one, as it depends on several experimental factors, such as flow-rate, temperature, particle size and concentration of the feed solution. As a matter of fact, retention on ion-exchangers depends on the distribution ratio of the ion on the resin, the stability constants of the complexes in solution, the exchange kinetics and the presence of other competing ions. Even though ion-exchangers recover hydrated ions, charged complexes and ions complexed by labile ligands, they are of limited use in practice for preconcentration of trace elements due to their lack of selectivity and their retention of major ions w17x. Yet, for some particular applications they may be a valuable tool. Hence, iron speciation was possible through selective retention of the negative Fe(III)-ferron complex on an anion-exchanger w18x. Selenium speciation was also feasible by selectively eluting Se(IV) and Se(VI) retained on a anion-exchanger w19x. 2.1.3. Elution of trace elements from the sorbent The same kind of interactions usually occur during the elution step. This time, the type of solvent must be correctly chosen to ensure stronger affinity of the trace element for the solvent, to V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 1182 Fig. 3. Disposable sorbent containers. ensure disruption of its interaction with the sorbent (as illustrated in Fig. 2). Thus, if retention on the sorbent is due to chelation, the solvent could contain a chelating reagent that rapidly forms a stronger complex with the trace metal. Elution may also be achieved using an acid that will disrupt the chelate and displace the free trace element. Similarly, if retention is due to ion exchange, its pH dependence enables the use of eluents with different pH to be used, such as acids. Of prime importance is to selectively elute only the target species. So, if they are more strongly retained on the sorbent than the interferent compounds, a washing step with a solvent of moderate elution strength is highly advisable before elution of the target species with the appropriate solvent. 2.2. Operation The sorbent may be packaged in different formats: filled micro-columns, cartridges, syringe barrels and discs w2,20,21x. The disposable sorbent containers are illustrated in Fig. 3. 2.2.1. Micro-columns The use of a micro-column is a common procedure for extraction of trace elements from various samples. It affords the opportunity of packing the column with the desired sorbent, so that a broader choice than the commercially disposable containers is available. In addition, the size of the column (i.e. the sorbent weight) may be adapted to the sample volume. In particular, it allows larger sample volumes, thus enabling the preconcentration of metal ions at very low concentration levels. However, such columns must be reused, so that careful blank washings should be conducted to avoid cross-contamination. In addition, columns with a narrow internal diameter limit usable flowrates to a range 1–10 mlymin that necessitates long trace-enrichment times for large sample volumes w22x. As will be discussed later, micro-columns are frequently used in systems affording the on-line coupling of SPE to analytical techniques. However, in that case, the size of the column is limited to achieve acceptable analytical performance. 2.2.2. Disposable cartridges and syringe barrels Nowadays, the most frequently used design in off-line SPE is the cartridge or the syringe barrel. They are usually made of polypropylene or polyethylene and filled with packing material having different functional groups. The solid sorbent is contained between two 20 mm polypropylene frits (in some cases they may be made of glass). They afford great selectivity due to the broad types of sorbents contained in commercially available systems with different column volume available. In addition, their disposable character prevents possible cross contamination. Cartridges vary from as little as 100 mg to 1 g or more. Syringe barrels range in size from 1 to 25 ml and packing weights from 50 mg to 10 g. Solvent reservoirs may be used at the top of the syringe barrels to increase the total volume (50– 100 ml). The barrel of the syringe terminates in a male Luer tip, which is the standard fitting to be used with various SPE vacuum manifolds available. For cartridges, both a female and male Luer tips are present, to enable use of either a positive or negative pressure. The major disadvantages of cartridges and syringe barrels are slow sample-processing rates and a low tolerance to blockage by particles and adsorbed matrix components, due to their small cross-sectional area. Channeling reduces the capacity of the cartridge to retain analytes and results V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 in contamination of the isolated analytes with impurities originating from the manufacturing and packing process. Such contaminants were evident for C18-silica cartridges, while less contaminants were observed with C18-silica disks w23,24x. 2.2.3. Disks The use of flat disks with a high cross-sectional area may largely prevent all the problems encountered with columns, cartridges and tubes w21x. The packing material is usually embedded in an inert matrix of polytetrafluoroethylene (PTFE) microfibrils, with a typical composition of 90% wyw sorbent and 10% wyw PTFE fibers w25x. Other types of disks use a glass-fibre matrix to hold the sorbent particles, in order to enable a higher flowrate. The disks are available in different diameters from 4 to 90 mm, the size most frequently used being 47 mm. They are designed to be used in conjunction with a filtration apparatus connected to a water aspirator w25x. In order to remove potential interferences and to ensure optimal extraction of the analyte of interest, disk cleaning and conditioning should be done before its use. Due to a lower void volume and a higher surface area associated with small particles as compared to cartridges, partitioning of the analytes is favored. Hence, a smaller mass of sorbent is required to process a similar volume of sample. Disks thus present the advantage of reducing solvent volumes for both the conditioning and elution steps. Additionally, the decreased back-pressure encountered with these devices enables the use of high flow-rates, and their wide bed minimizes the chance of plugging. In addition, new technology for embedding the stationary phase prevents channeling and improves mass transfer. As classical disks are dedicated to the SPE of large-volume samples, new systems have very recently emerged that enable the use of disks for small-volume samples: the extraction disk cartridge (the disk is placed in a syringe-barrel format), and the 96-well microtiter plate configuration w20,21,26x. Such systems are primarily dedicated to biological samples. One of the drawbacks of using disks is the decrease in the breakthrough volume (which is the volume that can be percolated without analyte losses). In addition, disks have lower capacity than 1183 cartridges, so that for real samples (e.g. high content of natural organic matter in river water) incomplete retention of the target metal species may result w27x. As a consequence, disks are recommended when there is a strong interaction between the analyte and the sorbent. 2.3. Advantages of the technique Classical liquid–liquid extractions of trace elements are usually time-consuming and labor-intensive. In addition, they require strict control of extraction conditions, such as temperature, pH and ionic strength. For all these reasons, several procedures tend to be replaced by SPE methods. This technique is attractive as it reduces consumption of and exposure to solvents, their disposal costs and extraction time w28x. It also allows the achievement of high recoveries w29x, along with possible elevated enrichment factors. However, as different results between synthetic and real samples may be observed w30x, recoveries should be estimated in both cases as far as possible. In addition, SPE can be interfaced on-line with analytical techniques, such as liquid chromatography (LC) or atomic absorption spectrometry (AAS). Its application for preconcentration of trace metals from different samples is also very convenient due to sorption of target species on the solid surface in a more stable chemical form than in solution. Finally, SPE affords a broader range of applications than LLE due to the large choice of solid sorbents. 2.3.1. Preconcentration LLE requires the use of large volumes of highpurity solvent, thereby affording limited preconcentration factors. The use of SPE enables the simultaneous preconcentration of trace elements and removal of interferences, and reduces the usage of organic solvents that are often toxic and may cause contamination. Upon elution of the retained compounds by a volume smaller than the sample volume, concentration of the extract can be easily achieved. Hence, concentration factors of up to 1000 may be attained. 2.3.2. Preservation and storage of the species SPE allows on-site pre-treatment, followed by simple storage and transportation of the pre-treated 1184 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 samples with stability of the retained metallic species for several days w21,24,31,32x. This point is crucial for the determination of trace elements, as the transport of the sample to the laboratory and its storage until analysis may induce problems, especially changes in the speciation. In addition, the space occupied by the solid sorbents is minimal and avoids storage of bulky containers and the manpower required to handle them. 2.3.3. High selectivity SPE offers the opportunity of selectively extracting and preconcentrating only the trace elements of interest, thereby avoiding the presence of major ions. This is crucial in some cases, such as with spectrophotometric detection, since the determination of heavy metals in surface waters may necessitate the removal of non-toxic metals, such as Fe or Zn, when they occur at high concentrations w33x. It may also be possible to selectively retain some particular species of a metal, thereby enabling speciation. For example, salen I modified C18-silica is quite selective towards Cu(II) w34x, while chemical binding of formylsalicylic acid on amino-silica gel affords selectivity towards Fe(III) w35x. This high selectivity may also be used to remove substances present in the sample that may hinder metal determination, such as lipid substances in the case of biological samples w36x. 2.3.4. Automation and possible on-line coupling to analysis techniques SPE can be easily automated, and several commercially available systems have been recently reviewed w26x. Home-made systems have also been reported w37x. In addition, SPE can be coupled online to analysis techniques. On-line procedures avoid sample manipulation between preconcentration and analysis steps, so that analyte losses and risk of contamination are minimized, allowing higher reproducibility w38x. In addition, all the sample volume is further analyzed, which enables smaller sample volume to be used. However, in the case of complex samples, off-line SPE should be preferred due to its greater flexibility, and the opportunity to analyze the same extract using various techniques. 2.3.4.1. On-line coupling to liquid chromatography. On-line systems mainly use a micro-column. The sorbent is chosen not only for its efficiency in trapping analytes, but also for its compatibility with the stationary phase packed into the chromatographic column. Indeed, it is highly recommended to use the same packing in the precolumn and the chromatographic column to prevent losses in efficacy upon analysis. For the case of two different sorbents being used, the retention of the analytes in the precolumn should be lower than in the analytical column to ensure band refocusing at the head of the chromatographic column. On-line systems with several detectors have been reported, such as ultraviolet (UV) detector w39x or inductively coupled plasma mass spectrometer (ICP– MS) w40x, with detection limits in the 0.05–50 mgyl range. Detection limits as low as 0.5 ngyl could even be achieved by detection at the maximum absorption wavelength using a photodiode array UV detector w41x. Additional coupling may be feasible, such as the on-line coupling of supercritical fluid extraction (SFE) with an on-line SPE–LC system w42x. The coupling of SPE to LC via flow injection has also been reported using cold vapour AAS (CV-AAS) as the detection, enabling enrichment factors approximately 850 w43x. 2.3.4.2. On-line coupling to atomic absorption spectrometry. Olsen et al. w44x and Fang et al. w45,46x were the first to describe an on-line flowinjection (FI) sorbent extraction preconcentration system for flame AAS (F-AAS) using microcolumns packed with a cation-exchanger. Later, they also proposed a system for on-line flowinjection sorbent extraction preconcentration with electrothermal vaporization AAS (ET-AAS) using lead as a model trace element w47x. Since then, numerous papers reported FI with on-line preconcentration followed by AAS, as exemplified by determination of Cu, Cr(VI) or Pb w48–50x. Selected applications are reported in Table 1. The sorbent should provide for rapid sorption and desorption of the analytes to be used in FI systems w65x. In addition, it should be provided for a high selectivity. In practice, C18-silica is very frequently used as organic solvents (such as methanol) can Table 1 Applications of SPE to FI on-line preconcentration systems Matrix Trace elements Inorganic sorbents Sea and waste Cd waters Eluent Analysis method Recovery (%) Loading time (volume) Preconcentration LOD factor (ngyl) Sampling Ref. frequency (hy1) None HNO3 – HCl ICP–AES 97.5–104 1 min 86 1100 40 w51x Thiourea ICP–AES 2.5% in HNO3 HCl Spectrophotometry 95.2– 103.3 2 min 62 4300 24 w51x 106.3 2 min — 0.016 nM — w52x Thiourea F-AAS 93.5–101 1 min — 660 60 w53x DPTHfunctionalizedSiO2 TSfunctionalizedSiO2 8-HQfunctionalizedSiO2 MBTfunctionalizedSiO2 Sea and waste waters Cd None Sea waters Fe None Geological sample, Cu metal, Pb nitrate Certified ore samples, Ni alloy, anode slime, electrolytic solution Fish, human urine Ag None Ag, Au, Pd None AmidinothioureidoSiO2 Thiourea F-AAS 98.7–101.4 1 min (4.5 ml) — 1100–17 000 — w54x MeHg, EtHg, PhHg, Hg(II) MeHg, Hg(II) Cu, Cd APDC C18-silica MeOHACNwater LC–CV-AAS 92–106 20 min (58.5 ml) 750–950 5.5–10.4 2.3 w43x DDTC C18-silica EtOH CV-AAS 85–107.5 (25 ml) 500 16 — w55x APDC C18-silica MeOH ET-AAS 6.5–1.26 — w56x EtOH F-AAS 104 s (0.5 ml) 30 s 25–100 C18-silica 86.7– 106.5 88.9– 100.5 22–32 300–6000 90 w57x MeOH ET-AAS 82.5– 111.2 110 s (0.5 ml) 25–50 1.7–4 9–12 w58x Sea water Certified sea waters Certified sea water, mussel, geological samples Certified sea waters Cu, Cd, Co 1,10Phenanthroline Cd C18-silica PAR or PADMAP V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Chelating Sorbent agent added 1185 1186 Table 1 (Continued) Trace elements Chelating Sorbent agent added Eluent Analysis method Recovery (%) Loading time (volume) Preconcentration LOD factor (ngyl) Sampling Ref. frequency (hy1) Sea water, industrial effluents Cr(III), Cr(VI), Cr(total) DDTC C18-silica MeOH F-AAS 95–105 60–300 s 90–500 20 30 w59x Sea, river waters Certified low alloy steel, mussel, tomatoe leaves Certified biological, vegetable samples Cr(VI), Cr(total) Co DDTC C18-silica EtOH ET-AAS 101–105 12 w60x C18-silica Acidified F-AAS EtOH 98–102 17.2 1600– 1800 3200 22 NN 1 min (3 ml) 30 s (3.25 ml) 90 w61x Pb DDTC C18-silica IBMK F-AAS 99.2– 137.9 2–10 min 60–189 3000 24 w62x Standard solutions Pb DDTP C18-silica EtOH F-AAS — 300–3000 — w63x Certified citrus leaves, marine sediment Standard solutions MeHg, Hg(II) DDTP C18-silica EtOH CV-AAS 99.6– 112.5 2.5–75 14–1000 min (10–150 ml) 4.5 min 20 (23.85 ml) 10 12 w64x Cu, Pb MeOH F-AAS — 80 s (4 ml) 14–60 4000– 10 000 — w65x Sea water Cu HCl ET-AAS 93.2–99.6 33 5 35 w66x Pharmaceutical preparations Sea water Zn DDTC, 8- C18-silica quinolinol or PAR None TAN-loadedC18-silica None TAN-loadedC18-silica None FZ-loadedC18-silica ZrO2 None HCl Spectrophotometry Spectrophotometry ICP–AES — 10 000 45 w67x 0.1– 0.3 nM 6–90 — w68x — w69x River, tap, rain waters Fe(II), Fe(III) Al, Bi, Cd, Co, Cr, Cu, Fe, Ga, In, Mn, Mo, Ni, MeOH HNO3 1 min (3 ml) 89.8–107.8 — — 95.4–99 2–20 min 6–60 (4–40 ml) 33 min 100 (100 ml) V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix Table 1 (Continued) Matrix Urine Eluent Analysis method Recovery (%) NHq ICP–AES 4 HNO3 or F-AAS NHq 4 HNO3 or F-AAS NHq 4 HNO3 ICP–AES Loading time (volume) Preconcentration LOD factor (ngyl) Sampling Ref. frequency (hy1) — 90–106 (2 ml) 35 s 10 25 200 800–1000 — 55 w70x w71x 86–117 — — — w72x )93 (10 ml) 50 42 000– 81 000 50 — w73x 1 min (12 ml) 3 min (37.8 ml) 340 50 40 w48x 80 800 18 w49x None None Acidic Al2O3 Acidic Al2O3 None Al2O3 None Basic Al2O3 Cu APDC PTFE turnings IBMK F-AAS 94–102 Cr(VI) APDC PTFE turnings IBMK F-AAS 95.5– 100.5 Pb APDC PTFE turnings IBMK F-AAS 95–102 3 min (39 ml) 330 800 15 w50x Cr(VI) APDC PTFE (KR) EtOH ET-AAS 105 19 4.2 21.2 w74x Cr(VI) APDC PTFE (KR) EtOH ET-AAS — 1 min (5 ml) 1 min (5 ml) 16.3 16 16.7 w75x Cd, Co, Cu, Dithizone PTFE (KR) Zn IBMK F-AAS 95.3– 108.4 1 min (5 ml) 23.4–69.3 1060– 2560 18 w76x Cd DDTC PTFE (KR) IBMK F-AAS 97.9–110 50 s (4.2 ml) 30 100 55 w77x Cr(VI) APDC PTFE beads EtOH ET-AAS 104–108 1 min (5 ml) 30.1 8.8 16.7 w75x 1187 Organic sorbents Tap, river, coastal waters Tap, river, coastal, industrial waste waters Tap, river and coastal waters, marine sediment, fish and mussel tissues Drinking, sea waters Certified natural water, sea water Certified human air, pig liver, sea prawn Certified humain hair and rice powder Certified natural water, sea water Pb, Tl, V, Sb, Sn, Zn Cr(VI) Cr(III), Cr(VI) Cr(III), Cr(VI) Cr(III) Chelating Sorbent agent added V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Waters Lake, river, tap waters Sewage waters Trace elements 1188 Table 1 (Continued) Trace elements Chelating Sorbent agent added Eluent Analysis method Recovery (%) Standard solutions Pb DDTP PUF EtOH F-AAS Certified seawaters Standard samples Alloys, ores Cu None None HClEtOH HCl F-AAS Cd, Zn Pd, Pt, Rh ODETA HClEtOH Cu, Cd, Ni Cu, Cd, Pb Cr(III) — PAN-coatedXAD-4 BSQ-loadedXAD-7 Highly crosslinked polystyrene IDA-Novarose Sulfasarzene None Cu, Cd, Ni, Zn, Mn Cd, Cu, Pb, Zn Cr(VI), Cr(total) Cd, Ni, Pb None Cu Cd Bathocuproine None Pb DDTP Tap water River, ground waters River, mineral and tap waters Estuarine waters Synthetic sea water Natural waters Certified pig kidney, rye grass, rice flour, tomato leaves Ground water Water, vegetable samples Standard solutions None DPC None Loading time (volume) Sampling Ref. frequency (hy1) 14–1000 300–3000 — w63x 29.1–296.1 60–600 — w32x 10 1600– 1900 3000– 8000 — w78x 30 w79x F-AAS 2.5–75 min (10–150 ml) 99.6–103 (2.5– 25 ml) — 20 s (0.1 ml) 77.8–103.6 1 min — HCl ICP–AES — 10 min 500–1000 — — w80x Lewatit TP807’84 PAPhA HCl 80–120 50 min (100 ml) 90 s (6.6 ml) 50 w33x 35 2000– 5500 200 — HCl Spectrophotometry F-AAS 30 w81x Toyopearl AFChelate 650 M Cation exchanger Cation exchanger Cation exchanger HNO3 ICP–MS 87–10 1 min (1 ml) — 1.4–86 — w82x HNO3 F-AAS — — 50–105 30–200 60 w45x HNO3acetone HCl Spectrophotometry ICP–MS 93–08 — — 8.9–15.2 — w83x 96.7–03.7 2 min (7.8 ml) 10 1000–4000 90 w84x HNO3 Spectrophotometry Spectrophotometry F-AAS — 10 min (8.3 ml) 90 s (5 ml) 2.5–75 min (10–150 ml) — 80 — w85x — 230 20 w86x 14–1000 300–3000 — w63x Anion exchanger Anion exchanger Activated carbon HNO3 EtOH Fluorimetry — Preconcentration LOD factor (ngyl) 97–01 94–104 — V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix Table 1 (Continued) Matrix Trace elements Eluent Analysis method Recovery (%) Loading time (volume) Preconcentration LOD factor (ngyl) Sampling Ref. frequency (hy1) Steels, Al solutions Bi DDTP EtOH ET-AAS 87–104.3 48 7 w87x Cu APDC IBMK F-AAS 88–113 4 min (10 ml) 2 min (4.5 ml) 14 Tap, mineral, well, river, swimming pool waters Silicon, tap water Rock, copper ore Activated carbon Activated carbon 100 600 17 w88x Fe None HCl MPT–AES 97.4–105 1000–36 000 — w89x None 1 min (1.2 ml) 3 min (15.6 ml) 4.3–6.4 Pd Activated carbon Activated carbon 145 300 w90x Thiourea F-AAS 103–107 15–20 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Chelating Sorbent agent added 1189 1190 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 be used as eluting solvents leading to a high sensitivity in flame AAS. Complexing reagents are, therefore, added for efficient retention of trace metals. Their choice is based on their fast reaction with metals, such as diethyldithiocarbamate (DDTC) and ammonium pyrrolidine dithiocarbamate (APDC) w56,59,60,62,91x. In addition, both reagents are water soluble and do not adsorb on C18-silica so that it does not overload with the reagent itself. However, these reagents lack selectivity, so that other reagents have been used for particular applications, like 1-10-phenanthroline w57x, 4-(2-pyridylazo)resorcinol (PAR) or 2-(2pyridylazo)-5-dimethylaminophenol (PADMAP) w58x, 0,0-diethyl-dithiophosphate (DDTP) w63x, lnitroso-2-naphthol (NN) w61x. l-(2-tiazolylazo)-2naphthol (TAN) w66x. The microcolumn can be inserted into the tip of the PTFE capillary in the autosampler arm of a graphite furnace atomic absorption spectrometer w56x. Even though C18-silica has been the most frequently used sorbent for FI preconcentration, other sorbents were found satisfactory for some applications as shown in Table 1, such as functionalized silica w51–53x, alumina w70–73x, activated carbon w63,87–90x, polyurethane foam (PUF) w63,92x, or PTFE turnings w48–50x. A particular knotted reactor (KR) has been recently developed, which consists of a long tube properly knotted usually made of PTFE. The trace element species are adsorbed on the inner wall of the tubing as indicated by scanning electron microscopy w77x. This reactor allows higher sample loading volumes than micro-columns due to its lower back-pressure. In addition, the inner wall may be precoated with a hydrophobic ligand for subsequent retention of trace elements w93x. However, on the other hand, lower enrichment factors are attained as compared to micro-columns w75x. For that reason, in case of trace metals, micro-columns are usually preferred for the achievement of low levels of determination. 2.3.4.3. On-line coupling to ICP–AES or ICP– MS. The first report of FI on-line preconcentration coupled to ICP-atomic emission spectrometry (AES) appeared nearly twenty years ago w94x. Since then, several studies have used this coupling with different sorbents such as ZrO2 or function- alized silica gel for example w51,69x. Similarly, numerous studies have reported on-line coupling to ICP–MS, as noted earlier w95x. A few examples are given in Table 1 w82,84x. 2.3.4.4. On-line coupling to spectrophotometry. Spectrophotometry offers the advantage of requiring inexpensive and very common instrumentation. In addition, by choosing a non-selective chromogenic reagent, multi-metal determinations may be possible w33x. Its coupling to FI analysis is well suited for monitoring purposes and a few studies present such systems as indicated in Table 1 w33,52,68,86x. Solid-phase spectrophotometry (SPS) has also been reported with FI systems due to its simplicity and low detection limits. The solid sorbent is packed in either commercially available or customized flow cells. With such systems the retained analytes are periodically removed from the flow cell using an acid or a complexing solution w67,83,85x. On-line FI sorbent extraction procedures have several advantages over the corresponding off-line methods: higher sample throughput (increased by 1 to 2 orders of magnitude), lower consumption of sample and reagent (also reduced by 1 to 2 orders of magnitude), better precision (with relative standard deviations approx. 1–2%), lower risk of loss or contamination and easy automation. However, the FI method by using column extraction has some disadvantages. In particular, it may suffer from insufficient adsorption on the resin and clogging of the column when insoluble ligands are used w96x. 3. Step-by-step method development guide Development of an SPE method can be considered as a two-step procedure. First, the most appropriate sorbent for the application should be chosen (the following is intended to help the reader in choosing a solid sorbent for trace element determination). Optimization of the most influential parameters should then be undertaken. Obviously, optimization should initially be performed using spiked synthetic solutions, but it must be followed by the use of certified reference materials or spiked real samples, as matrix components V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 (such as ligands or other ions) may change the trace element retention on the sorbent, thereby decreasing recoveries of the target species. 3.1. Selection of solid sorbent Solid sorbents may be hydrophobic or polar. It is common to call reversed-phase sorbents the packing materials that are more hydrophobic than the sample, which are frequently used with aqueous samples. On the other hand, normal-phase sorbents refer to materials more polar than the sample and they are used when the sample is an organic solvent containing the target compounds. When hydrophobic supports are used, retention of ionic metal species will require the formation of hydrophobic complexes. This can be achieved through addition of the proper reagent to the sample or thorough immobilization of the reagent on the hydrophobic solid sorbent. Addition of reagent to the sample is appropriate for the fixation of unstable metal species (such as Cu(I) and Fe(II)) to maintain speciation, while immobilization offers the convenience of having a prepared cartridge or disk before analysis. Immobilization may also provide a significant development in speciation analysis, because metal equilibrium in the sample may not be affected by reaction on the cartridge. The nature and properties of the sorbent are of prime importance for effective retention of metallic species. Careful choice of the sorbent is thus crucial to the development of SPE methodology. In practice, the main requirements for a solid sorbent are: (1) the possibility to extract a large number of trace elements over a wide pH range (along with selectivity towards major ions); (2) the fast and quantitative sorption and elution; (3) a high capacity; (4) regenerability; and (5) accessibility. In particular, sorbents that allow fast reaction rates are preferred to achieve faster extraction as well as higher loading capacities. Hence, sorbents based on hydrophilic macroporous polymers and cellulose or on fibrous materials provide excellent kinetic properties w97x. The broad variety of sorbents available explains one of the most powerful aspects of SPE, which is selectivity. Sorbents can be mainly categorized 1191 as organic based ones (natural polymers, as well as synthetic polymers) and inorganic based ones (silica gel SiO2, alumina A12O3, magnesia MgO and other oxide species). Immobilization of organic compounds on the surface of the solid support is usually aimed at modifying the surface with certain target functional groups for a higher selectivity of the extraction. The selectivity of the modified solid phases towards certain metal ions is attributed to several well-known factors, such as the size of the organic compound used to modify the sorbent, the activity of the loaded surface groups, and the type of the interacting functional group. However, the selective extraction of a single trace element from other interfering ion(s) represents a direct challenge for finding a suitable phase capable of exhibiting a sufficient affinity to selectively bind that metal ion. For particular applications, the combination of two sorbents may thus be advisable. As an example, the passage of water samples through two successive chelating resins enabled the determination of trace and major elements w98x. Similarly, the combination of an anion and a cation-exchange resin enabled the speciation of Cu and Mn in milk samples w99x. 3.1.1. Inorganic based sorbents Inorganic based sorbents are mainly made of silica gel even though other inorganic oxides may be used, as discussed later (cf. Fig. 4). Silica gel based sorbents present the advantages of mechanical, thermal and chemical stability under various conditions. They frequently offer a high selectivity towards a given metal ion. However, all silicabased sorbents suffer from different chemical limitations, namely the presence of residual surface silanol groups (even after an end-capping treatment) and a narrow pH stability range. Applications of such sorbents to off-line SPE are presented in Tables 2 and 3. 3.1.1.1. Silica gel. Silica gel can be used as a very successful adsorbing agent, as it does not swell or strain, has good mechanical strength and can undergo heat treatment. In addition, chelating agents can be easily loaded on silica gel with high stability, or be bound chemically to the support, affording a higher stability. 1192 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Fig. 4. Sorbents based on inorganic supports. The surface of silica gel is characterized by the presence of silanol groups, which are known to be weak ion-exchangers, causing low interaction, binding and extraction of ionic species w131x. In particular, silica gel presents high sorption capacity for metal ions, such as Cu, Ni, Co, Zn or Fe w132x. Retention is highly dependent on sample pH with quantitative retention requiring pH values over 7.5–8, as under acidic conditions silanol groups are protonated and the ion-exchange capacity of the silica gel is greatly reduced or even reduced to zero at low pHs. In addition, this sorbent has a very low selectivity, and is prone to hydrolysis at basic pH. Consequently, modification of the silica gel surface has been performed to obtain solid sorbents with greater selectivity. Two approaches are used for loading the surface with specific organic compounds, chemical immobilization and physical adsorption. In the first case, a chemical bond is formed between the silica gel surface groups and those of the organic compound (functionalized sorbent). In the second approach, the organic compound is directly adsorbed on the silanol groups of the silica gel surface (impregnated or loaded sorbent), either by passing the reagent solution through a column packed with the adsor- bent, or by soaking the adsorbent in the reagent solution. Impregnating reagents are ion-exchangers or chelating compounds. Numerous reagents have been investigated for impregnation of silica gel as a means of increasing retention capacity and selectivity of the sorbent for trace elements, namely thionalide (2-mercapto-N-2-naphthylacetamide) w101,102x, 2-mercaptobenzothiazole (MBT) w133x, NN w103x, 8-hydroxyquinoline (8-HQ) w134,135x, 3-methyl-l-phenyl-4-stearoyl-5-pyrazolone (MPSP) w100x, salicylaldoxime w132x, dimethylglyoxime (DMG) w13x, Aliquat 336 (methyltricaprylammonium chloride) and Calcon (hydrophobic sodium sulfonate) w136x. Examples of applications are given in Table 2. Increased stability of the sorbent is obtained by the chemical binding of chelating functional groups on silica gel w104x. Applications to the determination of trace elements have been reported for more than twenty years with several functional groups, such as amines, dithiocarbamates, iminodithiocarbamates or dithioacetals w13,105,106,137,138x. Careful choice of the bound chelating groups enables speciation studies. Hence, dithizone-functionalized silica gel was reported selective towards Hg(II) Table 2 Applications of off-line SPE to water samples using inorganic supports Matrix Silica Tap water Trace elements CuyCoyNi Operation Experimental conditions Analysis method Recovery (%) Adsorptive capacity Preconcentration factor LOD (mgyl) Ref. MPSP-loadedSiO2 Glass column Sample pH: 4.5 Elution: HCl 1 M Sample pH: 4 Washing elution: thiourea 0.2 MqHCl 0.1 M Sample pH: 7 Washing elution: NaBoratey NaOHy Iodine (pH 10) Sample pH: 3.5 Washing elution: acetic acid Elution: HCl 10 M F-AAS 94.6–101 43y45y49 mmolyg 40 60y40y70 w100x F-AAS 83–99 0.8 mgyg 3200 0.03 w101x Spectrophotometry 92–95 5.6 mmolyg — 0.12 w102x g-Emission 96–98 mmolyg 0.03 10–100 — w103x CV-AAS 99–99.5 300 mmolyg 200 3.96 w104x Sea water Pd Thionalideloaded-SiO2 Glass column (1 cm i.d) Sea water As(III) Thionalideloaded-SiO2 Glass column (1 cm i.d) River and sea water Co NN-loadedSiO2 Glass column Spiked tap water Hg(II) Column Tap and sea waters Hg(II) Spiked tap and sea waters Hg(II) DithizonefunctionalizedSiO2 DithioacetalsfunctionalizedSiO2 DithiocarbamatesfunctionalizedSiO2 Column Elution: water CV-AAS 91–100 917–1100 mmolyg 5 — w105x Glass column Elution: water CV-AAS 88–100 0.6–0.983 mmolyg — — w106x V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Sorbent 1193 1194 Table 2 (Continued) Trace elements Sorbent Operation Experimental conditions Analysis method Recovery (%) Adsorptive capacity Preconcentration factor LOD (mgyl) Ref. Sea water VyCoyNiy GayYyMoy CdyCryPry NdySmyEu yGdyTbyDy yMoyEry TmyYbyLuy WyU 8-HQfunctionalizedfluorinated metal alkoxide glass Column (6 mm i.d, 30 mm bed height) Washing sample pH: 5 Washing elution: HNO3 0.5 M backflush ICP–MS 85–116 — 10 0.00037– 2200 ngyl w107x Cr(III)yCr (VI) TiO2 78.4–99.2 8125y6983 mgyg 100 0.030y0.024 w108x CdyCoyCu yFeyMnyNi yPb TiO2 F-AAS 89–100 5000 mgyg 300 0.01– 0.04 w109x Tap, ground waters Cr(III)y Cr(VI) Neutral Al2O3 Column (1 cm i.d) ET-AAS 99–100 — 25 0.01 w110x Tap, ground waters Se(IV)y Se(VI) Acidic Al2O3 Teflon column (1 cm i.d) Sample pH: 2 or 8 Elution: HNO3 0.5 or 1 M Sample pH: 8 Elution: HNO3 1 M andyor EDTA 0.1 M Sample pH: 6.5–7 Elution: NH3 1 Mq HNO3 4 M Sample pH: 2–8 Elution: NH3 0.1 M and 4 M ET-AAS Natural, waste, sea waters Glass column (1 cm i.d) Glass column (1 cm i.d) ET-AAS 90–98 23.2y2.0 mgyg 16y100 0.049y 0.80 w111x Other oxides Rain, river, sea, tap waters V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix Table 3 Applications of off-line SPE to water samples using C18-silica based supports Matrix Reagent Operation Experimental conditions Analysis method Recovery (%) Adsorptive capacity Preconcentration factor LOD (mgyl) Ref. TBT None 93.5– 111.5 — 1000 — w24x TPhT None Fluorimetry 81–89 — 100 — w112x Sea water TPhT None Conditioning sample drying elution: acidified ethyl acetate Conditioning: MeOHqNaCl Sample Washing Air drying Elution: 10y4 M FlOH in MeOH backflush Washing sample Washing air drying elution: MeOH GC–ECD Spiked sea waters Cartridge or Empore disk (25 mm) Bond Elut cartridge Fluorescence (after addition of flavonol to the eluate) — — 250 — w113x ET-AAS 94–97 — 40–75 0.007y0.05 w114x ET-AAS 67–108 — 50–100 — w115x No reagent Sea water Addition of the reagent to the sample Sea water SeySb APDC Sea water CdyZnyCuy MnyFeyNiy Co 8-HQ Bond-Elut cartridge (40 mm) Glass column (1.4 cm i.d) Glass column (1.4 cm i.d) Sample pH: 1.2 Washing elution: MeOH Sample pH: 8.9 Washing (waterq oxine) elution: MeOH V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Trace elements 1195 1196 Table 3 (Continued) Matrix Trace elements Reagent Loaded reagent onto the sorbent River water Cu Neocuproine Experimental conditions Analysis method Recovery (%) Adsorptive capacity Preconcentration factor LOD (mgyl) Ref. Empore disks (47 mm) Washing conditioning: MeOH Sample pH: 5.0 Drying Spectrophotometry (454 nm) 99.7– 102.6 940 mg Cu2q 50–100 0.12 w116x Spectrophotometry (484 nm) — — 20–40 0.40–3.8 w27x — — 0.080 w117x 200 0.14y 0.16y 0.14 w31x CRM water (SLRS-3), lake, river, drinking waters Cu(I) Bathocuproine Bond Elut cartridge Empore disk Tap waters, well water Fe Bathophenanthroline Empore disks (47 mm) Synthetic seawaters MeHgy PhHgy Hg(II) Dithizone Sep Pak cartridge Rain, lake, river waters MonoBTy DiBTyTBT yMonoPhT yDiPhTyTP hT Cu Tropolone Sep-Pak cartridge Quinone derivative Empore disks (47 mm) Tap, well and river waters Elution: isopentyl alcohol Conditioning: MeOHqwater Sample pH: 4.3 Elution: MeOHwater 90:10 (vyv) Washing activation: MeOHqwater Sample pH: 4–7 Elution: EtOHq NaClO4 Sample pH: 4qEDTA 0.001 M Washing Elution: MeOH Conditioning: MeOH sample pH: 2–3 Air drying Elution: diethyl ether Washing Conditioning: buffer sample pH: 7.0 Drying Elution: HNO3 0.1 M Spectrophotometry (533 nm) LC–DAD 95–104 Ethylation-GC– FPD — — — — w118x F-AAS 98.4–102 360 mg Cu2q 400 0.2 w119x V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Operation Table 3 (Continued) Trace elements Reagent Operation Experimental conditions Analysis method Recovery (%) Adsorptive capacity Preconcentration factor LOD (mgyl) Ref. Tap, rain, snow and sea waters Cu(II) Schiff’s base (salen I) Empore disk (47 mm) AAS — 396 mg Cu2q )500 0.004 w34x Synthetic and spring waters Pb(II) Schiff’s base Empore disk F-AAS 97.1– 100.2 700 mgydisk 50 16.7 w120x River water Pb(II) BAS Empore disk (47 mm) F-AAS — 476 mg Pb2q G300 0.050 w121x Sea waters Fe(II) Ferrozine Sep Pak cartridge Spectrophotometry (562 nm) 91 — 40 0.6 nmolyl w122x Rain, sea waters Fe(II) Ferrozine Sep Pak cartridge LC–UV (254 nm) 92–99 — 100–500 0.1 nmolyl w123x Certified sea waters (NASS-2 and SLEW-1) CuyCd APDC Teflon cartridge (0.94 mm i.d) ET-AAS 95.8– 103.3 — 25–50 0.0024y 0.00018 w124x Tap and Spring waters U(IV) TOPO Empore disk (47 mm i.d) Washing sample pH: 5.5–6 Air drying elution: HNO3 0.1 M Washing sample pH: 2–8 Air drying Elution: HNO3 O.5 M Washing sample pH: 2–7 Air drying elution: acetic acid 1 M Conditioning: MeOHqwater sample pH: 6.8–8.3 Washing Elution: MeOH Conditioning: MeOHqwater Sample Washing Elution: MeOH Conditioning: MeOHqwater Sample: 6–8 Air drying Elution: MeOH Washing conditioning sample elution: MeOH Spectrophotometry 85 4033 mgydisk 8 0.1 w125x V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix 1197 1198 Table 3 (Continued) Trace elements Reagent Operation Experimental conditions Analysis method Recovery (%) Adsorptive capacity Preconcentration factor LOD (mgyl) Ref. Sea waters LayCeyPry NdySmyEuy GdyTby DyyHoy EryTmy YbyLu Bi HDEHPy H2ME HP Sep Pak cartridge Sample pH:3–3.5 Washing elution: HCl 6 M ICP–MS 88.8–99.8 — 200–1000 — w126x Cyanex 301 Cartridge (0.5 g) ET-AAS 98.5–100 — 10–100 105 w127x Sea, well and tap waters Be Quinalizarin Sep pak cartridge F-AAS 98–101 200 mg 200 0.2 w128x Tap, spring waters Agq HT18C6 Empore disk AAS 996– 100.3 210 mg Agydisk 200 0.050 w129x Tap, river, well and spring waters Hg(II) HT18C6TO Empore disk (47 mm i.d) Conditioning: HCl 0.1 M Sample pH: 1 Air drying Elution: HNO3 3 M Washing conditioning sample pH: 6–6.6 Elution: HNO3 0.5 M Conditioning: MeOHqwater Sample elution: Na2S2O3 0.1 M Washing sample pH:-7 Air drying Elution: HBr 1 M CV-AAS 97.1– 101.3 241 mg Hg2q 50 0.006 w130x Spiked natural waters V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 w104x, even though dithizone was reported to react with many trace elements w139x. Similarly, purpurogallin-bound silica gel enabled selective extraction of Fe(III) w140x. Simultaneous retention of trace elements is possible by choosing a nonselective chelating group, such as N-propylsalicylaldimine w141x or Bismuthol I (2,5-dimercaptol,3,4-thiadiazole) w142x. Acidic groups can also be used for further chelation of trace elements, such as phosphonic acid w143x and calixarene tetrahydroxamic acid w144x. Alternatively, macrocycles may be bound to silica (SGBM) w145x, such as 18-crown-6 (18C6) w146x. It must be kept in mind that despite chemical bonding of functional groups on the silica gel surface, free silanol groups still remain w143x. Their number can be minimised by end-capping the sorbent, but some will still be present. As a consequence, they will participate in the retention of trace elements somewhat, especially at pHs above their pKa (ionized form). 3.1.1.2. C18-bonded silica gel. Despite the large variety of bonded phases available, octadecylbonded silica has currently become the most popular phase used. Numerous applications report the use of C18-silica, as indicated by the studies reported for water samples in Table 3. In particular, organometallic compounds (e.g. tributyltin (TBT), triphenyltin (TPhT), alkylselenides) can be retained on this sorbent due to possible hydrophobic interaction w19,24,112,113x. Bare C18-silica can also retain a fraction of inorganic trace elements, probably due to the presence of silanol groups on its surface w34x. However, in practice, due to its hydrophobic character, C18-silica is not well suited for retention of trace element species, as the latter are often polar or ionic. Retention on C18-silica may be improved by addition of a ligand reagent to the sample before its percolation through the sorbent. The hydrophobic part of the ligand will thus have hydrophobic interaction with the C18silica and be retained on the sorbent, while the functional group of the ligand will ensure chelation of the trace element. Among reagents, one can cite 8-HQ w115x, APDC w114x, 1,10-phenanthroline w18x, or bathocuproine w27x. 1199 An alternative approach is to form the complex by passing the sample through a C18-silica containing the immobilized reagent. Octadecyl bonded silica, modified by suitable ligands has been successfully used for the separation and sensitive determination of metal ions. Examples are given in Table 3. The careful choice of the ligand may add selectivity to the extraction step, favoring speciation. For example, salen I-modified C18silica was found selective for Cu(II) w34x, while impregnation of C18-silica with neocuproine was suitable for Cu(I) w116x. C18-silica { coated with bis wl-hydroxy-9,10-anthraquinone-2-methylxsulfide (BAS) was preferred for Pb(II) retention w121x, while coating with N,N9-diethyl-N9-benzoylthiourea (DEBT) was recommended for Pd w147x. Macrocycles may also be loaded on C18-silica and efficiently used for the retention of trace metals, such as hexathia-18-crown-6 (HT18C6) w129x or calixarene hydroxamate w144x. For some particular applications, mixed ligand complexes may be used to ensure synergistic adsorption of the metal complex on the solid sorbent. Thus, while Cu(II) ions cannot complex with neutral tri-n-butyl phosphate (TBP) molecules adsorbed on C18-silica, the form of Cu(TTA) complex (TTA being 2-thenoyltrifluoroacetone) was retained at approximately 80% w148x. Alternatively, in some cases, loading the chelating agent on C8-silica instead of C18-silica may give better results as observed for the retention of bismuth on oxinate-loaded reversed phase w149x. Despite their broad application to trace element preconcentration, bonded silica phases (either C18silica or functionalized-silica gel) present the drawback of a limited range of pH that can be used, as in acidic (below 2 to 4) and basic (above 8) pHs hydrolysis may occur, which changes the interactions that occur between the sorbent and the trace elements. As a consequence, polymeric sorbents may be preferred. 3.1.1.3. Other inorganic oxides. Apart from silica other inorganic oxides have been tested for the adsorption of trace elements as shown in Tables 1 and 2. Whereas SiO2, due to its acidic properties, is expected to adsorb only cations, basic oxides (such as magnesia MgO) should adsorb only 1200 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Fig. 5. Sorbents based on organic supports. anions. As a matter of fact adsorption of ions on oxide surfaces is believed to proceed with participation of hydroxyl groups. These groups are negatively charged (deprotonated) under basic conditions, thereby retaining cations and positively charged (protonated) under acidic conditions, thereby retaining anions. Consequently, on amphoteric oxides (namely titania TiO2, alumina Al2O3, zirconia ZrO2), cations are adsorbed under basic conditions (pH above the isoelectric point of the oxide which was reported to be 6.2 for TiO2 w109x) while anions are adsorbed under acidic conditions (pH below the isoelectric point of the oxide). For example, chromium speciation may be achieved by careful adjustment of the sample pH: pH 2 and 7 for retention of Cr(VI) (anionic) and Cr(III) (cationic), respectively, on acidic alumina w70,71x; pH 2 and 8 for retention of Cr(VI) and Cr(III) on titania, respectively w108x. The concurrent adsorption of Hq is responsible for the absence of retention of cationic species at very low pHs. However, changing the sample pH may affect speciation and should be avoided as far as possible. So it may be preferred to find a suitable sorbent for retaining the targeted species with subsequent selective elution for further speciation studies. With regards to chromium speciation, neutral alumina has been used for that purpose w110,150x. The preparation technique is of prime importance w109x, as the adsorption properties of many oxides strongly depend on the characteristics of the solid, namely crystal structure, morphology, defects, specific surface area, hydroxyl coverage, surface impurities and modifiers. Thus, the coating of acidic alumina with an anionic surfactant allowed the selective retention of Cr(VI) in very acidic solutions w151x. Adsorption on inorganic oxides may also be influenced by the presence of salts in the matrix. In particular, high concentrations of phosphates and sulfates may decrease trace element retention on titania w152x. On the opposite, major cations (Naq, Kq, Ca2q, Mg2q) are weakly adsorbed on titania w152,153x. 3.1.2. Organic based sorbents Organic based sorbents may be divided into polymeric and non-polymeric sorbents, as shown in Fig. 5. Polymeric sorbents have been, by far, the most used for trace element preconcentration having the advantage over bonded silica in that they can be used over the entire pH range. Their V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 disadvantage is that the conditioning step is more time consuming as they require extensive cleaning before use. Comprehensive reviews on polymeric phases have been published w97,154x. Masque´ et al. published an extensive review on sorbents used for the SPE of polar organic micropollutants from natural waters w155x. The purpose of this section is to summarize the most frequently used organic based sorbents for trace elements, as well as the more recently reported ones. In most applications, new sorbents have been synthesized by chemically bonding chelating groups to polymeric cross-linked chains and characterizing their ability to selectively adsorb trace elements. Most of the chelating groups reported have low water solubility to avoid their leaching from the sorbent, as most applications deal with aqueous samples. At the same time, a too hydrophobic group will hinder wettability of the sorbent by the aqueous sample, resulting in poor retention efficiency. A compromise is thus necessary. In addition to the functional group, the efficiency of polymeric sorbents depends on various physicochemical parameters, such as particle size, surface area, pore diameter, pore volume, degree of crosslinking and particle size distribution. 3.1.2.1. Polystyrene-divinylbenzcne based sorbents. Macroporous hydrophobic resins of the Amberlite XAD series are good supports for developing chelating matrices. Amberlite XAD-1, XAD-2, XAD-4 and XAD-16 are polystyrenedivinylbenzene (PS-DVB) resins with a high hydrophobic character and no ion-exchange capacity. In addition to the hydrophobic interaction that also occurs with C18-silica, such sorbents allow p–p interactions with aromatic analytes. Due to the hydrophobic character of PS-DVB, retention of trace elements on such sorbents requires the addition of a ligand to the sample. Inorganic ligands may be used w156x, but organic ligands are preferred, such as APDC w157,158x, 8HQ w159,160x, or diphenylcarbazide (DPC) w161x. Alternatively, ligands may be attached to the PSDVB by physical adsorption such as dithizone w162x, PDT (3-(2-pyridyl)-5,6-diphenyl-l,2,4-triazine) w4x, tropolone w163,164x, l-(2-pyridylazo)2naphthol (PAN) w32,165x, DDQ (7-dodecenyl- 1201 8-quinolinol) w166x, APDC w167x or 5-BrPA-DAP (2-(5-bromo-2-pyridylazo)-5-(diethylamino) phenol) w168x. Macrocycles can also be adsorbed, such as calixarene hydroxamate w144x. However, in practice, the resins prepared by impregnation of the ligand are difficult to reuse, due to partial leaching of the ligand (thus resulting in poor repeatability). To overcome this problem, the resin may be chemically functionalized. Several chemical modifications of PS-DVB have recently been reviewed w169x, but only a few are commercially available. The ligands are generally coupled to a methylene or an azo spacer on the matrix. Among {ligands, one can cite Alizarin Red-S w170x, salicylic acid (SA) w171x, thiosalicylic acid (TSA) w172x, pyrocatechol violet (PV) w173x, chromotropic acid (CA) w174x, pyrocatechol (PC) w175,176x, Tiron (disodium salt of l,2-dihydroxybenzene-3,5-disulfonic acid) w177x, quinalizarin (1,2,5,8-tetrahydroxy-anthraquinone) w22x, bicine wN,N-bis (2-hydroxy-ethyl) glycinex w178x, and poly(dithiocarbamate) (PDTC) w179x. Of great interest are also the sulfonated PSDVB resins, as they show excellent hydrophilicity and high extraction efficiencies for polar organic compounds w154,155x. In the case of rapid sulfonation under mild conditions, a mixed-mode retention can be observed: adsorption of neutral compounds on the polymeric resin, and cation exchange of ionic species on sulfonate groups w180x. The use of a particular sulfonated PS-DVB resin has been recently reported to enable chromium speciation w181x. 2-Naphthol-3,6-disulfonic acid (NDSA) has been coupled to the PS-DVB through an azo function. In that way, the formation of an azo cation at very low pHs enabled retention of the anionic Cr(VI), whereas the sulfonate group enabled retention of Cr(III) in neutral and basic media w181x. For particular applications, trimethylammonium functionalized PS-DVB may also be used with anion-exchange properties. As summarized in Table 4, which presents selected applications of polymeric sorbents for the preconcentration of trace elements from water samples, even though PS-DVB has been probably the most widely used of polymers, others also have been successfully used as detailed below. 1202 Table 4 Applications of off-line SPE using polymeric sorbents to water samples Matrix Trace elements Adsorptive resins Tap water CdyCuy MnyNiy PbyZnq 8-HQ Operation XAD-2 Polypropylene Conditioning column sample pH: 8–9 Elution: HCl 2 M Glass column Washing (1 cm i.d) sample pH: 1 Elution: H2SO4 0.05 MyMeOH Glass column Washing (0.9 cm i.d) conditioning sample pH: 6 Elution: acetone Column Washing (4 mm i.d) sample pH: 8 Elution: EtOHy HNO3 2 M Syringe barrel Purification sample pH: 11 Elution: HCl 2 M Tap water Cr(VI), total CrqDPC XAD-16 Drinking and sea waters BiyCdyCoy CuyFeyNiy PbqAPD C Chromosorb102 Tap, mineral, river waters Coq8-HQ Chromosorb105 Tape, lake, waste waters Cr(III) Cellulose Chelating resins Waters TBT TropoloneloadedXAD-2 Glass column (1.5 cm i.d) Experimental conditions Sampleq 0.8% H2SO4 Washing Elution: IBMK Analysis method Recovery (%) Adsorptive Preconcentration LOD capacity factor (mgyl) Ref. ICP–AES 82.3– 97.2 — F-AAS 100 — w160x 97.3–99.0 0.4 mgyg 5.25 45 w161x F-AAS 95–110 300 0.10–11 w158x ET-AAS 95.2–99.2 — 80 0.0134 w182x ET-AAS 98–99.3 — 100 0.0018 w183x ET-AAS 104 — 80 0.0144 w164x — V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Sorbent Table 4 (Continued) Trace elements Sorbent Operation Experimental conditions Analysis method Recovery (%) Adsorptive Preconcentration LOD capacity factor (mgyl) Ref. Sea water CoyCuy FeyNiy Zn PDT-loadedXAD-2 Column (0.9 cm i.d) F-AAS 96.3– 103.5 — — — w4x Well water, river water CuyCdy CoyPby ZnyMn QuinalizarinfunctionalizedXAD-2 Glass column (1 cm i.d) Washing elution: MeOH in Soxhlet apparatus Washing sample pH: 5–7 Elution: HNO3 4 M F-AAS 91–98 100y50y40y50y 100y65y65 2.0y1.3y5.0y w22x 15.0y1.0y1.6 Well waters ZnyCdy PbyNi PVfunctionalizedXAD-2 Glass column (1 cm i.d) F-AAS 98 60y50y 23y18 — w173x Well waters ZnyPb Salicylic acidfunctionalizedXAD-2 Glass column (1 cm i.d) F-AAS 98–100 3.15y 1.70y 1.62y 5.28y 1.42y 0.94y 2.19 mgyg 1410y 1270y 620y 1360 mgyg 1146y461 mgyg 180y140 — w171x Well waters ZnyCdy PbyNi Alizarin Red-SfunctionalizedXAD-2 Glass column (1 cm i.d) F-AAS 95–100 511y124y 306y124 mgyg 40 10 w170x River waters CuyCdyCo TironyNiyPbyZny functionalizedMnyFe XAD-2 F-AAS 91–99 — 25–200 0.5–24 w177x Glass column (1 cm i.d) Washing sample pH: 3–7 Elution: HNO3 4 M Washing sample pH: 5.0 Elution: HCl 1 M-2–4 M Washing sample pH: 4–6 Elution: HNO3 1–4 M or HCl 4 M Washing sample pH: 4–7.5 Elution: HNO3 4 M V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix 1203 1204 Table 4 (Continued) Trace elements Sorbent Operation Experimental conditions Analysis method Recovery (%) Adsorptive Preconcentration LOD capacity factor (mgyl) Ref. River waters CdyCoyCu yNiyFeyZn CAfunctionalizedXAD-2 Glass column (1 cm i.d) Washing sample pH: 4–7 Elution: HNO3 or HCl 2 M F-AAS 95–100 River and tap waters Pb CAfunctionalizedXAD-2 Glass column (1 cm i.d) F-AAS 97 River and tap waters CdyCoyCu yFeyNiyZn PCfunctionalizedXAD-2 Glass column (1 cm i.d) F-AAS — River and tap waters Pb PCfunctionalizedXAD-2 Glass column (1 cm i.d) F-AAS River and tap waters Pb TSAfunctionalizedXAD-2 Glass column (1 cm i.d) Tap, river waters CdyCoyCuy FeyNiyZn TSAfunctionalizedXAD-2 Glass column (1 cm i.d) Washing sample pH: 3–8 Elution: HNO3 2–10 M Washing sample pH: 3–6.5 Elution: HNO3 2 M Washing sample pH: 5–7.5 Elution: HNO3 1 M Washing sample pH: 4 Elution: HNO3 0.5–2 M Washing sample pH: 3.5–7 Washing elution: HNO3 2 M 9.35y 3.84y 8.50y 3.24y 6.07y 9.65 mgyg 186.3 mmolyg 100–200 — w174x 200 4.06 w176x 0.023– 0.092 mmolyg 80–200 — w175x 94 104.7 mmolyg 100 3.80 w176x F-AAS 93 89.3 mmolyg 100 4.87 w176x F-AAS 92–98 197.5y 106.9y 214.0y 66.2y 309.9y 47.4 mmolyg 180–400 0.48y0.20y 4.05y0.98y 1.28y3.94 w172x V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix Table 4 (Continued) Matrix Trace elements Artificial sea water, natural waters Sorbent Experimental conditions Analysis method Recovery (%) Adsorptive Preconcentration LOD capacity factor (mgyl) CdyCuyMny APDC-loadedNiyPbyZn XAD-4 Glass column (0.9 cm i.d) Sample pH: 5.0 Washing elution: HNO3 4 M ICP–AES Artificial sea water, natural waters CdyCuy MnyNiy PbyZn pipDTC-loadedXAD-4 Glass column (0.9 cm i.d) Sample pH: 5.0 Washing elution: HNO3 4 M ICP–AES Sea water AgyAly BiyCdy CuyFey GayMny NiyPbyTi DDQ-loadedXAD-4 F-AAS or ETAAS Tap water CuyMnyZn Calixarene TetrahydroxamateloadedXAD-4 F-AAS — Tap and mineral waters Mn PDTCfunctionalizedXAD-4 F-AAS 97.2 Spiked solutions CoyCuy FeyHgy NiyPbyZn BicinefunctionalizedXAD-4 Teflon column Washing (8 mm i.d) sample pH: 8 Washing elution: HCl 2 M Backflush Plastic Sample cartridge pH: 8.5 (0.9 cm i.d) Elution: acidified water (pH 2.0) Glass column Conditioning (1 cm i.d) sample pH: 10 Elution: HNO3 8 M Glass column Conditioning sample pH: 5.5–7 Elution: HCl 1 M 98.2–99.6 9.47y 11.08y 8.62y 7.21y 10.25y 10.62 mgyg 97.6–99.1 9.18y 10.76y 8.17y 7.46y 9.86y 10.28 mgyg 73–107 0.55 mmolyg ET-AAS 97.6–99.1 0.32–0.44 mmolyg Ref. 180y230y120y 130y160y215 0.1y0.4y0.3y w167x 0.4y0.6y0.5 150y200y140y 120y150y200 0.7y1.0y0.8y w167x 0.9y1.7y1.2 62.5 0.00016–0.3 w166x — 25 — w144x 9.1 mmolyg 20 0.5 w179x 40–50 — w178x V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Operation 1205 1206 Table 4 (Continued) Trace elements Sorbent Operation Experimental conditions Analysis method Recovery (%) Waste waters Cr(III)yCr (VI) NDSAfunctionalizedPS-DVB Glass column (1 cm i.d) F-AAS 85.9–96.1 0.40y1.18 mmolyg 20 — w181x River water Hg(II) PAAfunctionalizedPS-DVB Glass column (1 cm i.d) Spectrophotometry 96 0.6 mmolyg 30 — w184x River and tap waters Pb XO-loadedXAD-7 Glass column (1 cm i.d) F-AAS 91 16.9 mmolyg 100 2.44 w176x River waters CdyCoyCu yFeyNiyZn XO-loadedXAD-7 Glass column F-AAS 96–100 1.6–2.6 mgyg 10–200 9y24y6y 6y3y21 w185x River and reservoir waters Cr(III) 8-HQfunctionalizedpolyacrylonitrile Glass column (4 mm i.d) ICP–MS 98–105 41.7 mmolyg 5 0.06 w186x Aqueous AuyPty sample PdyIr from a nonferrous metal smelter Aminothioureafunctionalizedpolyacrylonitrile Glass column (0.5 cm i.d) ICP–AES 97–99 2.80y 1.75y 1.56y 1.15 mmolyg 6–65 — w187x Sea water Aminophosphonic- Glass column dithiocarbamate(4 mm i.d) functionalizedpolyacrylonitrile Sample pH: 1.5 or 6 Elution: HCl 4 M Conditioning sample pH: 5.4 Elution: H2SO4 2 M Washing Sample pH: 5 Elution: HNO3 1 M Sample pH: 4–5 Elution: HCl 1 or 2 M Washing sample pH: 6 Washing Elution: HCl 2 My HNO3 0.1 M Washing sample pH: 2 Washing elution: HCl 4 My CS(NH2)2 3% Sample pH: 6 Washing elution: HCl 2 M ICP–MS 93–104 0.83–74.1 mmolyg 200 0.002–0.601 w188x ngyl BeyBiy CoyGay AgyPby CdyCuy MnyIn Adsorptive Preconcentration LOD capacity factor (mgyl) Ref. V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix Table 4 (Continued) Trace elements Sorbent Operation Experimental conditions Analysis method Recovery (%) Adsorptive Preconcentration LOD capacity factor (mgyl) Ref. River, lake and rain waters Hg(II)y MeHg Dithiocarbamatefunctionalizedpolyvinyle Column (1.5 cm i.d) CV-AAS 91–95 — 667 0.0002 w189x Sea water CdyCuy MnyNiy PbyZn Chelamine Column ET-AAS 91–102 1 mmolyg 200 0.0023– 0.033 w190x Sea water CdyCoy CuyMny NiyPbyZn Chelex-100 Column ET-AAS or F-AAS )98 — 50 0.001–0.1 w191x River water, seawater CdyPbyZn Amberlite IRC-718 Glass column (0.5 cm i.d) F-AAS 63–104 1.06y 0.096y 1.77 mmolyg 10 — w192x Ground and fresh waters Se(IV)ySe (VI)y SeCyst Amberlite IRA-743 Glass column (1 cm i.d) Washing Sample pH: 1–11 Elution: thiourea 5% in HCl Washing sample pH: 6.5 Washing elution: HNO3 2 M Sample pH: 6.5 Washing elution: HNO3 2 M Washing conditioning sample washing elution: HNO3 Washing conditioning sample elution: HClO4 1 Mqwater LC–ICP– MS 93–97 — 55 0.010 w193x Anion exchanger (SAX) Cartridge (0.5 g) — — 8 w194x Ion exchangers Waste water Cr Activation: MeOHq buffer Sample pH: 4.5 Elution: Na2SO4 0.5 M Spectro80–98.8 photometry (544 nm) after reaction DPC V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix 1207 1208 Table 4 (Continued) Trace elements Sorbent Operation Experimental conditions Analysis method Recovery (%) Adsorptive Preconcentration LOD capacity factor (mgyl) Ref. Sea, river, tap waters Se(IV)y Se(VI) Anion exchanger (SAX) Cartridge GC–MS 91–99 — 40–500 1600y 1400 w19x River, spring, waste waters Be Anion exchanger Column (1 cm i.d) Conditioning Sample pH: 7 Elution: HCOOH 1 MqHCl 3 M Conditioning: HClqH2Oq NaOH Sample pH: 6–8 Elution: HCl 1.5 M F-AAS 95–102.5 — 125 0.045 w195x V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Matrix V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 3.1.2.2. Divinylbenzene-vinylpyrrolidone copolymers. Sorbents made of divinylbenzene-vinylpyrrolidone (DVB-VP) copolymers have recently been developed, such as Oasis HLB w154x. The hydrophilic N-vinylpyrrolidone { affords good wettability of the resin, while the hydrophobic divinylbenzene provides reversed-phase retention of analytes. This sorbent has been successfully applied to the determination of polar organic compounds in water samples. It is more convenient to use, compared to classical sorbents, as it can dry out during the extraction procedure without reducing its ability to retain analytes. In addition, it is stable over the entire pH range. However, until now, no application related to the preconcentration of trace elements has been reported. Similarly, the use of Oasis MCX, a sulfonated divinylbenzenevinylpyrrolidone copolymer w154x, may be useful for the retention of trace element species, as this sorbent combines the properties of the previous sorbent with those of a strong cation-exchanger. Application to the preconcentration of triphenyltin has been reported w36x. 3.1.2.3. Polyacrylate polymers. Amberlite XAD-7 and XAD-8 are ethylene-dimethacrylate resins. They are non-aromatic in character and possess very low ion-exchange capacity. Due to the polarity of acrylates, such resins enable the recovery of polar compounds. However, this polarity is quite moderate so that most of the time reagents are added to increase retention. Direct addition to the sample is sometimes performed. As an example, Cu(II) forms a complex with 8-hydroxyquinoline5-sulfonic acid (8-HQ-5-SA), which can be further retained on Amberlite XAD-8 as an ion-pair with cetyltrimethylammonium (CTA) w15x. Yet, most of the time chelating reagents have been loaded on such resins, mainly Amberlite XAD-7, to increase their retention capacity for trace elements andyor their selectivity such as 8-(benzenesulfonamido)quinoline (BSQ) w78x, xylenol orange (XO) w176,185x, 5-BrPADAP w168x, or dimethylglyoxal bis(4-phenyl-3-thiosemicarbazone) (DMBS) w196x. Such loaded sorbents are stable for several months and can be reused. For a higher stability, chemical binding of the chelating group may be performed, as reported for dithizone on 1209 poly(ethylene glycol dimethacrylate-hydroxyethylmethacrylate) microbeads w197x. 3.1.2.4. Polyurethane polymers. Due to its sorption capacity for several trace elements polyurethane foam has been tested for use in SPE. Most of the time complexing reagents are added to enhance the sorption capacity. Hence, PUF coated with DMG, NN, DDTC or hexamethylenedithiocarbamate (HMDC) was found efficient in retaining trace elements w198–200x. The chelating reagent can also be directly added to the sample, and the metal chelates further retained on PUF, as observed with thiocyanate complexes w92,201,202x and DDTP complexes w63x. Very recently, the immobilization of an enzyme (alkaline phosphatase) has been reported on PUF with further application as an enzymatic procedure for Pb(II) determination w203x. 3.1.2.5. Polyethylene polymers. Polyethylene is also attractive for SPE of trace elements as this support adsorbs several metal complexed with hydrophobic ligands. Additionally, the adsorbed complexes can be eluted with a small volume of organic solvents permitting high enrichment factors. Polyethylene can also be used in strongly acidic and basic media and for that reason, it has been used as a sorbent for the retention of chromium in an acidic medium after the addition of DPC w204x. 3.1.2.6. Polytetrafluoroethylene polymers. PTFE can retain trace elements after addition of a chelating reagent to the sample such as APDC, DDTC or dithizone (DZ) w48–50,74,76x. PTFE may also be precoated with a suitable ligand, like 2-methyl8-hydroxyquinoline w93x. The sorbent can be used as PTFE turnings w48–50x, PTFE beads w75x, or as a PTFE tubing in a knotted reactor w74,76x. 3.1.2.7. Polystyrene polymers. Polystyrene polymers may be an interesting alternative to common sorbents (namely Amberlites XAD-2 and XAD-8, C18-silica) when they have a hyper cross-link structure. The addition of a reagent to the sample is required to form complexes that are further retained on the hydrophobic sorbent w79x. 1210 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 3.1.2.8. Polyamide polymers. Polyamide polymers have been used for the retention of rare earth elements w205x. A chelating reagent was added to the sample for complexing the trace elements. This reagent Thorin (o-w3,6-disulfo-2-hydroxy-lnaphthylazoxbenzenearsonic acid) was chosen to enable interaction with the sorbent through electrostatic forces and non-hydrophobic interaction. 3.1.2.9. Iminodiacetate-type chelating resins. Polymeric resins containing iminodiacetate groups w–CH2–N(CH2COOy)2 x as active sites (IDA resins) have been widely used for the retention of trace elements. They have been synthesized by bonding the iminodiacetate functional groups to several polymeric sorbents, such as polystyrene (Chelex-100) w10,29,30,191,206–209x or a highly crosslinked agarose gel (IDA-Novarose) w80x. The spacer arm length was found to have an effect on the formation of metal complex species in the chelating resin w12x. A major drawback of such sorbents is that, due to the weak acid character of the functional group, the degree of protonation will critically affect the ability of the resin to retain metal cations. Hence, for Chelex-100, protonation of the carboxylates and the donor N atom are reported to be complete at pH 2.21, while a completely deprotonated form is reached at pH 12.30. Also, such sorbents are non-selective, so that trace element retention may be reduced due to retention of major ions (namely Ca(II) and Mg(II)) w99,191x. Besides, the presence of ligands in the sample may prevent trace element retention on the sorbent due to their complexation as observed in real waters due to the presence of organic matter w80,210x. 3.1.2.10. Propylenediaminetetraacetate-type chelating resins. The synthesis of a fine-particle macroporous polymer-based propylenediaminetetraacetic acid (PDATA) type resin has been recently reported w211x. The structure of this sorbent is very similar to that of ethylene diamine tetraacetic acid (EDTA) with a spacer arm enabling the retention of several trace elements upon chelation. 3.1.2.11. Polyacrylonitrile based resins. Polyacrylonitrile fibers have been functionalized to obtain ion-exchange chelating sorbents with aminophosphonic, dithiocarbamate or aminothiourea groups w187,188x. However, as such synthesis are timeconsuming an alternative is to coat the polyacrylonitrile fiber with a proper reagent for further trace element retention such as 8-HQ w186,212x. 3.1.2.12. Ring-opening metathesis polymerisationbased polymers. A high-capacity carboxylic acidfunctionalized resin has been prepared using ring-opening metathesis polymerisation (ROMP) w213x. Electron microscopy revealed that the obtained material consists of irregularly shaped, agglomerated particles having a non-porous structure with diameter and specific surface area dependent on the polymerisation sequence and the stoichiometries. This material was pH stable and could be reused. The presence of the carboxylic groups confers an excellent hydrophilic character to the sorbent (ensuring a high wettability of the sorbent by water), while the polyunsaturation of the carrier chain, as well as the entire backbone provides for a significant reversed-phase character. The carboxylic acid groups provide weak coordination sites enabling the retention of rare earth elements. Similarly, dipyridyl amide-functionalized resins have been reported to allow the extraction of ‘soft’ metals such as Pd(II) and Hg(II) w214x. 3.1.2.13. Carbon sorbents. Activated carbon is prepared by low-temperature oxidation of vegetable charcoals. Due to their large surface areas (300–1000 m2 yg), these sorbents are well-recognized for their very strong sorption both for trace organic compounds and trace elements. There is evidence of two types of adsorption sites on activated carbons: (1) graphite-like basal planes that enable adsorption through van der Waals forces, especially p-electron interactions, and (2) polar groups like carbonyls, hydroxyls and carboxyls, that may interact via ionic interaction of hydrogen bonding w215x. Consequently, trace elements may be directly adsorbed on activated carbon w19,216x. Metal chelates may also be retained on this sorbent after addition of a proper chelating agent to the sample w17x such as amino acids w217x, dithizone w218x, APDC w88,219x, PAN w220x, 8-HQ w221x, cupferron w221x, Bismuthiol II V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 (3-phenyl-5-mercapto-l,3,4-thiadiazole-2(3H)-thione) w222x, or DDTP w63,87x. The ligand should be chosen to avoid a strong interaction with the activated carbon otherwise complete dissociation of the metal chelate would be observed w215x. The main drawback when using activated carbons is their heterogeneous surface with active functional groups that often lead to low reproducibility. In addition, these sorbents are very reactive and can act as catalysts for oxidation and other chemical reactions. Fortunately, along with the development of polymer materials and bonded phases, a new generation of carbon sorbents appeared in the 1970s and 1980s with a more homogeneous structure and more reproducible properties. Graphitized carbon blacks (GCB) are obtained from heating carbon blacks at 2700–3000 8C in an inert atmosphere w155x. They are nonspecific and non-porous sorbents (surface area approx. 100 m2 yg), and are considered to be both reversed-phase sorbents and anion-exchangers due to the presence of positively charged chemical heterogeneities on their surface. Such sorbents have been extensively used in the past few years for the SPE of polar organic pollutants from water samples w155x, but their use for trace elements is still rare w223x. Their main drawbacks are possible irreversible retention of analytes, which may be overcome by elution in the backflush mode, and poor mechanical stability. Porous graphitized carbon (PGC) is a more stable carbon based sorbent than GCB as the graphite is immobilized on a silica substrate. So this sorbent should be suitable for trace element retention even though until now its applications have been limited to trace organic compounds. 3.1.2.14. Cellulose. Cellulose was found effective in retaining trace elements present in water samples either directly or upon addition of a chelating agent to the sample w183,224x. In particular, the selective retention of Cr(III) was reported thereby enabling chromium speciation w183x. This sorbent may also be functionalized to increase the SPE selectivity. Thus, selenium speciation has been reported on cellulose functionalized with quaternary amine due to the selective elution of the 1211 retained Se(IV) and Se(VI) species using nitric acid at two different concentrations w225x. 3.1.2.15. Naphthalene based sorbents. Retention of trace elements on microcrystalline naphthalene is also feasible, either after addition of a ligand to the sample w226x, or after functionalization of the solid to ensure better adsorption characteristics towards trace elements w227,228x. However, the use of this solid support is rather uncommon. In addition, until now it has been reserved to batch experiments. 3.2. Influential parameters The main experimental variables that affect analyte recovery by SPE have been extensively reported by Poole et al. w2,229x. They are briefly discussed below and illustrated with reported applications. 3.2.1. Conditioning parameters 3.2.1.1. Washing step. A washing step is highly recommended, especially when ultratraces of elements are to be determined. Thus, blank extracts containing trace levels of Zn, Cu and Fe were suspected to be due to contaminants from C18silica w115x. 3.2.1.2. Conditioning solvent. Even though some sorbents have been used without a conditioning step this is not recommended. This step will at least remove possible remaining contaminants and air from the sorbent bed. Additionally, in some cases, this step is crucial for successful retention of the analytes. The nature of the conditioning solvent must be appropriate to the nature of the solid sorbent to ensure good wettability of the functional groups. As an example with hydrophobic supports such as C18-silica or PS-DVB, quite polar organic solvents such as methanol should be used. The sorbent should further be conditioned by a solvent whose nature is similar to that of the sample. Thus, for aqueous samples, the solvent will be water with a pH and ionic strength similar to that of the sample. 1212 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 require lower maximum flow-rates than disks ranging typically from 0.5 to 5 mlymin. This value may be increased by a factor of 10 using disks. Fig. 6. Typical representation of the breakthrough curve (i.e. concentration of the analyte at the outlet of the SPE system vs. sample volume percolated through the system). VB is the breakthrough volume, VR the chromatographic elution volume and VC the sample volume corresponding to the retention of the maximum amount of analyte, C0 is the initial analyte concentration in the sample. 3.2.2. Loading parameters 3.2.2.1. Sample volume to be percolated. An important parameter to control in SPE is the breakthrough volume, which is the maximum sample volume that should be percolated through a given mass of sorbent after which analytes start to elute from the sorbent resulting in non-quantitative recoveries (Fig. 6). The breakthrough volume is strongly correlated to the chromatographic retention of the analyte on the same sorbent and depends on the nature of both the sorbent and the trace element, as well as on the mass of sorbent considered and the analyte concentration in the sample w3x. In addition, it depends on the sorbent containers, as disks usually offer higher breakthrough volumes than cartridges. This volume may be determined experimentally or estimated using several methods w229x. For that purpose the nature of the sample has to be taken into account, as the possible presence of ligands may dramatically reduce the breakthrough volume w230x. 3.2.2.2. Sample flow-rate. The sample flow-rate should be optimized to ensure quantitative retention along with minimization of the time required for sample processing. This parameter may have a direct effect on the breakthrough volume, and elevated flow-rates may reduce the breakthrough volume w4,229x. As a rule, cartridges and columns 3.2.2.3. Sample pH. Sample pH is of prime importance for efficient retention of the trace elements on the sorbent. Its influence strongly depends on the nature of the sorbent used. Careful optimization of this parameter is thus crucial to ensure quantitative retention of the trace elements and in some cases selective retention. In particular with ionexchangers, correct adjustment of sample pH is required to ensure preconcentration. Thus, in the case of cationic-exchangers, low pH usually results in poor extraction due to competition between protons and cationic species for retention on the sorbent. When retention of trace elements is based on chelation (either in the sample or on the solid sorbent), the sample pH is also a very important factor as most chelating ligands are conjugated bases of weak acid groups and accordingly, they have a very strong affinity for hydrogen ions. The pH will determine the values of the conditional stability constants of the metal complexes. By contrast, pH may have no influence with some non-ionizable organic ligands w130x. For inorganic oxides, pH is also of prime importance. In particular, on amphoteric oxides such as TiO2 or Al2O3, cations are adsorbed at elevated pHs due to the deprotonation of functional groups, whereas anion retention requires acidic conditions for the protonation of functional groups. 3.2.2.4. Sample matrix. The presence of ligands in the sample matrix may affect trace element retention when stable complexes are formed in the sample with these ligands, as trace elements are less available for further retention. Thus, if metals are present in the sample as strong complexes, they may not dissociate resulting in no retention of the free metal on the sorbent. As an example, reduction in the retention of Cu(II) on Amberlite CG50 occurs in the presence of ligands such as glycine w230x. In the case of real samples, the presence of natural organic matter is of great concern as it may complex trace elements as observed for Cu(II) w231,232x. Yet, in some cases V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 the presence of ligands may be a valuable tool for adding selectivity to the SPE step. This requires that the added ligands be correctly chosen to complex only the elements that are not of interest, so that they are not retained on the sorbent w31x. The presence of ions other than the target ones in the sample may also cause problems during the SPE step. In particular, due to their usually high levels (e.g. Ca(II)), they may hinder the preconcentration step by overloading the sorbent or cause interferences during spectrophotometric analysis. Therefore, their influence should be studied before validating a SPE method. Sometimes the addition of a proper masking agent (such as EDTA, thiourea or ethanolamine for example) may prevent the formation of interferences due to ions present in the sample w128x. Finally, the ionic strength of the sample is another parameter to control for an efficient SPE, as it may influence the retention of trace elements, and thus the value of the breakthrough volume for a given sorbent w122,195x. 3.2.3. Elution parameters 3.2.3.1. Nature of the solvent. The nature of the elution solvent is of prime importance and should optimally meet three criteria: efficiency, selectivity and compatibility, as discussed below. In addition, it may be desirable to recover the analytes in a small volume of solvent to ensure a significant enrichment factor. The eluent may be an organic solvent (when reversed-phase sorbents are used), an acid (usually with ion-exchangers), or a complexing agent. Firstly, the eluting solvent should be carefully chosen to ensure efficient recovery of the retained target species and quantitative recovery as far as possible. As an example among several solvents tested for the elution of TPhT from C18-silica (namely a Triton X-100 surfactant aqueous solution, acetonitrile, tetrahydrofuran (THF), methanol–water 80:20, methanol), only methanol enabled the achievement of acceptable recoveries (approx. 85%) w112x. A further characteristic of the elution solvent arises with the possibility of introducing selectivity. Using a solvent with a low or moderate eluting 1213 power, the less retained analytes can be recovered without eluting the strongly retained compounds. Thus, if the elements of interest are those that remain on the sorbent another elution step with a more eluent solvent will ensure their quantitative recovery. In that way interferent analytes were removed during the first eluting step (also called washing step). On the opposite, if the compounds of interest are the less retained on the sorbent their elution with a low or moderate eluting solvent ensures their selective recovery, as the interferent compounds will remain on the sorbent due to stronger interactions with the solid support. In some cases, this selectivity may authorized speciation. For example, 1 M HCOOH removed only Se(IV) from an anion-exchange resin, leaving Se(VI) retained on the sorbent, which was further eluted using 2 M HCl w19x. Finally, the elution solvent should be compatible with the analysis technique. In particular, when using both flame and electrothermal AAS, HNO3 should be preferred to other acids (namely H2SO4, HCl), as nitrate ion is a more acceptable matrix w128x. 3.2.3.2. Solvent pH. As retention of trace elements on solid sorbents is usually pH-dependent, careful choice of the elution solvent pH may enhance selectivity in the SPE procedure. As an example, once retained on eriochrome black-T (ERT)-functionalized-silica gel, Mg(II) could be eluted first at a pH approximately 4, while increasing the pH to 5–6 was required for eluting Zn(II) w9x. 3.2.3.3. Elution mode. Most of the time, for practical reasons, sample loading and elution steps are performed in a similar manner. However, to avoid irreversible adsorption and ensure quantitative recoveries, elution in the backflush mode is recommended in some cases. This means that the eluent is pumped through the sorbent in the opposite direction to that of the sample during the preconcentration step. This is especially crucial when carbon-based sorbents have to be used due to possible irreversible adsorption of the analytes. 3.2.3.4. Solvent flow-rate. The flow-rate of the elution solvent should be high enough to avoid 1214 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 excessive duration, and low enough to ensure quantitative recovery of the target species. Typical flow-rates are in the range of 0.5 to 5 mlymin for cartridges and of 1 to 20 mlymin for disks w34x. As a rule, the higher the flow-rate, the larger the solvent volume required for complete elution w119,121,129,130x. 3.2.3.5. Solvent volume. Similar to the breakthrough volume, the elution volume may be determined either experimentally or estimated theoretically w229x. Minimum elution volume for a cartridge is defined as 2 bed volumes of elution solvent. Bed volume is typically 120 mly100 mg of sorbent. For classical disks, the minimum solvent volume required is approximately 10 mlymg of sorbent w20x. Consequently, larger elution volumes would be required for disks. The elution volume can usually be reduced by increasing the concentration of the eluting solvent (e.g. acid). However, in this case, problems with subsequent analysis may be encountered (e.g. F-AAS). Alternatively, the use of micro-sized disks may allow reduced solvent volume w20x. The elution step should enable sufficient time and elution volume to permit the metallic species to diffuse out of the solid sorbent pores. As a rule, 2 elution cycles are usually recommended as compared to a single step (e.g. 2=5 ml elution should be preferred to a single 10 ml elution). Soaking time is also critical and 2 to 5 min soak is most of the time allowed before each elution. 4. Applications of SPE to the determination of selected trace elements 4.1. Cadmium Cadmium is known to be a highly toxic trace metal. Owing to its very low concentrations in the environment, a preconcentration is usually required for its determination. This can be performed on an anion-exchange resin, after reaction of Cd(II) with chloride and the formation of the anionic CdCl2y 4 complex w86x. Yet, Cd(II) retention generally occurs through chelation, either by adding a chelating agent to the sample, by impregnation of the sorbent, or by the synthesis of new chelating resins. Hence, a FI preconcentration on-line with F-AAS has been reported for the determination of Cd(II) w77x. Cadmium complexed with DDTC was sorbed on the inner walls of a PTFE knotted reactor, and further on-line eluted with isobutyl methyl ketone (IBMK) giving a detection limit of 0.1 mgyl. However, very acidic pHs (lower than 2) were required for optimum sensitivity and collection efficiency. In addition, DDTC is rather non-selective, so that retention of other trace elements may occur. This drawback maybe overcome by adding masking agents, such as thiourea and ascorbic acidyphenanthroline for copper and iron, respectively, or by choosing a more selective reagent, such as DDTP w233x. Still very acidic pHs were required for optimum sensitivity and collection efficiency. APDC offers the advantage of a broader pH range without decomposition. Hence, Cd(II) complexed with APDC was stable for pHs between 4 and 8, and could be efficiently retained on C18silica w56,124x. Further elution with methanol and FI on-line ET-AAS analysis enabled detection limits of 0.178 ngyl or 1.26 ngyl depending on the system used. Other chelating reagents may be used, such as PAR or PADMAP w58x, or tetra-(4bromophenyl)-porphyrin (T4BPP)w41x. Impregnation of the sorbent with a chelating reagent has been reported for the preconcentration of Cd(II). In that case, the choice of a chelatant that have a high affinity for cadmium is preferred to ensure high selectivity. For example, BSQ has been immobilized on Amberlite XAD-7 and used in a FI system giving a detection limit of 1.9 mgy l w78x. However, Zn(II) was also retained, while other ions were found to interfere (namely Mg(II), Cu(II), Fe(III)). Amberlite XAD-7 coated with DMBS also enabled the retention of Cd(II) from neutral medium with simultaneous retention of Pb(II) w196x. When ICP–AES is used as the analysis technique, the use of organic solvents should be avoided as they may generate strong turbulence in the ICP. So, new chelating resins were developed for the Cd(II) preconcentation in a FI–ICP–AES system, such as l,5-bis(di-2-pyridyl) methylene dithiocarbohydrazide (DPTH)- or methylthiosalicylate (TS)-functionalized silica gel leading to detection limits of 1.1 and 4.3 mgyl, respectively V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 w51x. They offer the advantage of having no affinity for sodium, potassium, calcium and magnesium, enabling the Cd(II) analysis in real water samples. However, other trace elements reduced Cd(II) retention on both resins, such as Zn(II) or Cu(II). In some cases, the interfering effect of Zn(II) can be avoided by careful adjustment of the sample pH, as observed on the Lewatit TP807’84 resin that contains a phosphonic derivative as extractant w33x. Nevertheless, other ions were still co-extracted with Cd(II), namely Cu(II) and Pb(II). 4.2. Chromium Chromium species enter the environment as a result of effluent discharge from steel industries, electroplating, tanning industries, oxidative dyeing, chemical industries and cooling water towers. They may also enter drinking water supply systems from the corrosion inhibitors used in water pipes and containers or by contamination of the underground water from sanitary landfill leaching. Therefore, it is of major concern to study the characteristics of chromium in aquatic systems. Chromium occurs mainly in (III) and (VI) oxidation states. While Cr(III) is an essential trace element, Cr(VI) is highly carcinogenic and mutagenic due to its high oxidative character. So, it is important to develop analytical methods that enable analysis of chromium in its different oxidation states. However, sampling and preconcentration steps might disturb the redox equilibrium between Cr(III) and Cr(VI), thereby affecting the original speciation state of the sample. Most SPE methods are based on the high reactivity of Cr(VI), due to the relatively inert nature of Cr(III). Many methods are thus based on the determination of Cr(VI) and total chromium. FI on-line preconcentration procedures reported for chromium species were exhaustively reviewed up to 1992 by Sperling et al. w71x, and later (until 1998) by Prasada Rao et al. w59x. 4.2.1. Cr(VI) The determination of chromium is frequently achieved by spectrophotometry after derivatisation with a reagent such as DPC. The reaction of DPC 1215 is very selective for Cr(VI) so it can be performed directly without a separation step. The chromate oxidizes DPC to diphenylcarbazone (DPCO) to form a soluble strongly red–violet compound with Cr(III) (Cr(III)-DPCO(3yn)q ). A large excess of DPC is essential as compounds present in the sample may consume the reagent. The Cr(III)DPCO complex can be retained on polyethylene packed in a column and subsequently eluted with methanol before being analyzed by LC w204x. This procedure was applied to the determination of chromium in geological samples. The determination of total chromium may also be achieved after preliminary oxidation of Cr(III). Elements posing interference with this method are mainly Mo(VI), Fe(III) and V(V), and their separation using SPE with an anionic-exchanger (SAX) has been performed for the subsequent spectrophotometric determination of chromium w194x. The Cr(III)DPCO complex was also found to be retained on cation-exchange (SCX) membrane disks w234x. The color intensity of the membrane was then correlated to the Cr(VI) concentrations by visual analysis. This simple procedure was found to provide a highly sensitive semi-quantitative field test for the determination of Cr(VI) in aqueous samples. A cation-exchange resin was also used for the FI on-line preconcentration of the Cr(III)DPCO complex w83x, which can be retained at pH 1 on Amberlite XAD-16 resin and then eluted using 0.05 M H2SO4 solution in methanol w161x. In this way, Cr(VI) can be determined in tap water samples, as total chromium after oxidation of Cr(III) into Cr(VI) using potassium permanganate. Cr(VI) may also be chelated by DDTC, and subsequently retained on C18-silica w60x. Total chromium may be estimated using the same procedure with prior oxidation of Cr(III) to Cr(VI). The FI on-line preconcentration systems enabled detection limits of 16 ngyl for Cr(VI) and 18 ngy l for total Cr. Very recently, selective retention of Cr(VI), compared to Cr(III) was obtained using PTFE turnings packed in the micro-column of a FI manifold. APDC was used for complexation of Cr(VI) in the samples before preconcentration, and elution was achieved with IBMK before analysis using F-AAS w49x. Similarly, after reaction with APDC, the Cr(VI) complex could be retained 1216 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 on the PTFE inner wall of a knotted reactor or on PTFE beads packed into a column, and subsequently eluted with ethanol w74,75x. Also, acidic alumina enables selective retention of Cr(VI) w70x and a micro-column was used in a FI manifold to separate and preconcentrate Cr(VI) from Cr(III) in water samples. However, it was later reported that between pH 3 and 6 Cr(III) could also be retained on that sorbent w235x. 4.2.2. Cr(III) Very recently, selective preconcentration of Cr(III) on a cellulose micro-column has been reported to be highly dependent on sample pH w183x. Below pH 8, very low retention occurred, while at pH 11 almost quantitative adsorption could be achieved. Elution of the retained Cr(III) species with 5 ml of HCl (2 M) enabled its subsequent determination by ET-AAS with a 1.8 ngyl limit of detection. This method also enabled the determination of total chromium following initial reduction of Cr(VI) to Cr(III) with hydroxylamine. The coating of the positively charged acidic alumina with an anionic surfactant, SDS, has been reported to enable selective retention of Cr(VI) in strongly acidic solution, while the cationic Cr(III) remains unabsorbed w151x. Separation of the two species could thus be obtained and Cr(III) determined by ET-AAS. However, this method presents two major limitations; no analyte enrichment and the need to adjust sample pH to 0.6, which may affect chromium speciation. In another study, basic alumina has been shown to selectively retain Cr(III) at pH 2–7 w73x, permitting elution with nitric acid. Selective retention of Cr(III) was also reported with a macroporous PS-DVB resin (CHP-20P) after complexation in solution with 8-HQ at 85 8C w159x. The use of hydroxylamine as a reductant was found efficient for the reduction of Cr(VI) to Cr(III) without affecting Cr(III) retention as it does not complex this species. 8-HQ-immobilizedpolyacrylonitrile fiber has been recently reported to selectively retain Cr(III) at pH 6, enabling its preconcentration from river and reservoir water samples, while Cr(VI) was unretained w186x. Selective retention of Cr(III) (cationic) was also achieved on a chelating ion-exchange column packed with poly(aminophosphonic acid) (PaPhA) resin w81x. By using on-line FI preconcentration coupled to F-AAS, a detection limit of 0.2 mgyl could be obtained with a sample throughput of 30 hy1. 4.2.3. Cr(VI)yCr(III) A procedure based on the reaction of chromium species with APDC and subsequent retention of the complexes on SPE permitted a subsequent online LC–UV analysis w39x. The Cr(III) reacts with the chelating agent under relatively mild conditions to give only one product, i.e. triswpyrrolidine-ldithioato-S,S9xchromium(III), whereas Cr(VI) reacts to give two products, one being the former complex. Thus, the concentration of Cr(VI) needs to be corrected for the Cr(III) complex. The automated SPE system was optimized to yield detection limits of 0.2 mgyl for Cr(III) and 0.06 mgyl for Cr(VI). A FI on-line preconcentrarion procedure followed by F-AAS detection has been reported to allow determination of both chromium species in water samples based on selective formation of DDTC complexes of Cr(VI) in the pH range of 1–2, and of Cr(III) in the 4–9 pH range w59x. A detection limit of 0.02 mgyl could thus be achieved. Cr(III) could also be retained in the presence of Mn(II), which enhances the Cr(III) signal w59x. TiO2 has been recently reported to be very promising for chromium speciation w108x. Indeed, this sorbent can selectively adsorb Cr(III) or Cr(VI) depending on the pH of the sample. At pH 2, Cr(VI) is the sole chromium species retained, while at pH 8 it is Cr(III). Nitric acid (0.5 or 1 M) allows quantitative elution of the retained species. Acidic alumina has also been reported to retain both Cr(III) and Cr(VI) w235x. However, careful adjustment of the pH was of prime importance as below pH 3 and above pH 6 retention of Cr(III) and Cr(VI), respectively, decreased w235x. Hence, a FI on-line preconcentration method has been reported on activated acidic alumina with buffering of the sample before retention either at pH 7 (for Cr(III)) or pH 2 (for Cr(VI)) w71x. Cr(III) exhibited a typical cationic sorption (it increased with pH and decreased when competing V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 cations are present), whereas Cr(VI) exhibited a typical anionic sorption (it decreased with increasing pH and in the presence of competing dissolved anions). Thus, the use of neutral alumina seems more appropriate for chromium speciation since it offers the advantage of requiring no adjustment of the water sample pH. Hence, both Cr(III) and Cr(VI) can be quantitatively retained on neutral alumina with speciation being achieved using selective elution of the species, i.e. 1 M ammonia solution for Cr(VI) followed by 4 M HNO3 for Cr(III) w110x. This method affords a 25-fold preconcentration factor with a limit of detection of 10 ngyl. A similar preconcentration on alumina followed by selective elution for separation of the two chromium species has been reported w72x. A new resin that consists of PS-DVB functionalized with NDSA enabled the retention of Cr(VI) at sample pH of 1.5, and of Cr(III) at pH 6.5 w181x. Thus, speciation of chromium was accomplished by adjustment of the pH along with the percolation of two sub-samples at the desired pH. The selective retention of both chromium species could also be achieved on a polymeric sorbent containing aminocarboxylic groups at pH 3 and 7, for Cr(VI) and Cr(III), respectively w236x. Microwave heating was found to promote the sorption. Chromium species may also be retained on an anionic-exchanger (SAX) after their complexation with EDTA w237x. Controlled elution of the analytes with 0.5 M NaCl enables their speciation. In this manner, detection limits of 0.4 and 1.1 mgyl were obtained for Cr(III) and Cr(VI), respectively. 4.3. Iron Iron is widely distributed in nature and is one of the most important elements in biological systems. Its biological effectiveness is influenced by its chemical properties, such as valence, solubility and the degree of chelation or complex formation. Several methods have been proposed for the determination of Fe(III) and Fe(II) species. 4.3.1. Fe(III) Some solid phases have been synthesized to enable high selectivity towards Fe(III). Thus, the chemical bonding of formylsalicylic acid on ami- 1217 no-silica gel enabled the preconcentration of Fe(III) from a mixture containing other trace elements in batch experiments w35x. The capacity of this sorbent for Fe(III) was 0.95 mmolyg. This high selectivity was attributed to the presence of two chelating oxygen atoms. The selective SPE of Fe(III) on purpurogallin chemically immobilized on silica gel has also been reported w140x. Once eluted from the sorbent the iron species were analyzed using AAS. Due to strong affinity of Fe(III) to the bound organic compound, as compared to Fe(II) (the distribution coefficient, Kd, was found to be 120 500 for Fe(III) and 12 700 for Fe(II)), speciation of iron could be achieved using this procedure. Only minor Fe(III) (2.1%) was found to be reduced to Fe(II) upon interaction with the sorbent. Using batch experiments, Fe(III) could be selectively extracted from tap water as well as from a soft drink sample (7-Up). Yamini and Amiri w117x developed an efficient method for the selective extraction, concentration and determination of trace amounts of Fe(III) in aqueous media, enabling determination of iron in a 500 ml water sample in less than 30 min with a reproducibility better than 3%. Iron(III) was initially reduced to Fe(II) by addition of hydroxylamine (NH2OH) and the bathophenanthroline complex Fe(bathophenanthroline)2q was analyzed 3 by the use of C18-silica membrane disks and spectrophotometry. The ligand was added to the sample before SPE, and the solution heated to enhance formation of the complex. By comparison with AAS and ICP emission spectrometry, this method offers simplicity and applicability to field determination of iron. However, pure solvents eluted only a fraction of the complex and addition of NaClO4 to the elution solvent was required for complete removal of the retained complex, indicating that interactions between the complex and the C18-silica are dispersive or ionic. At sample pH below 2.5 recovery decreased probably due to competition of Hq with Fe(II) for reaction with bathophenanthroline. The influence of several ions was investigated, most were tolerated at high levels without interfering with the determination of iron. However, some species such as Co2q, Ni2q, 2q VOy , interfered. The inter3 and especially Cu 2q ference effect of Cu is due to the preferential 1218 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 formation of a very stable, almost colorless, complex between Cu(I) and bathophenanthroline. Its effect was eliminated by the addition of excess amounts of thiourea, as a masking agent, while the effect of the other ions was eliminated by use of excess amounts of bathophenanthroline. A FI catalytic spectrophotometric method has also been developed for the shipboard determination of iron in sea water samples w52x. Retention was achieved using 8-HQ-functionalized silica gel. After elution with HCl mixing with the reagents (H2O2 and N,N-dimethyl-p-phenylenediamine (DPD)) was performed to ensure the formation of the color. A detection limit of 0.016 nM could thus be achieved. 4.3.2. Fe(II) Iron(II) is thermodynamically unstable in sea water containing dissolved oxygen due to its rapid oxidation to Fe(III) (half-life being approx. 4–10 min depending on the pH). Yet, the determination of Fe(II) in sea water is important because of its role in the solubility, speciation and biological utilization of iron in oceanic surface waters; in addition Fe(II) may reduce oxygen, thereby producing radicals in the water. The colorimetric reagent ferrozine (FZ) w3(2-pyridyl)-5,6-bis(4-phenylsulfonic acid)l,2,4-triazinex forms a stable, colored complex with Fe(II) in a pH range of 4–10, but not with Fe(III) w238x. Interferences are also possible due to complexation of Cu(I), and to a lesser extent of Co(II) and Ni(II) by FZ. Thus, a procedure has been developed enabling the preconcentration of Fe(II) from sea water, thanks to its retention as the Fe(FZ)3 complex on a C18-silica cartridge and subsequent elution with methanol w122x. Iron was then analyzed directly by spectrophotometry (562 nm). C18-silica was loaded with FZ by passing a FZ solution through the cartridge (the retention capacity of the cartridge for FZ increased with increasing ionic strength of the solution). Once the sea water has been passed through the cartridge, washing was performed with 5 ml of 0.1 M NaCl–0.005 M NaHCO3 to remove sea salts and prevent precipitation of Mg2qand Ca2q in methanol upon elution. Cu(I) interference was minimised by adding neocuproine (NCH) to the methanol extract avoiding the presence of the Cu(FZ)2 complex. However, in this method both the Fe(FZ)3 complex and the excess FZ were eluted and contributed to the absorbance measurement, which may result in unreliable determination of low Fe(II) concentrations. Therefore, this procedure was further improved by performing a chromatographic separation of the two species after the SPE step. In addition, 254 nm was found more suitable for detection than 562 nm. In this way, ultratrace amounts of Fe(II) could be determined in several samples (aerosols, rainwater and sea water) w123x. A procedure, similar to that reported by King et al. w122x has been used for the on-line FI preconcentration of Fe(II) from sea water samples w68x. It could be extended to Fe(III) determination based on its initial reduction by addition of ascorbic acid to the sample. Copper interference could be suppressed by loading the C18-silica with a mixture of FZ and NCH, but at the same time the column capacity was lower than when loaded with FZ alone. 4.3.3. Fe(III)yFe(II) Most of the methods report the determination of only one of the two iron species mainly by selective complexation. However, this can cause a shift in the Fe(II)yFe(III) equilibrium in the solution as a result of redox reactions. To prevent such problems, a procedure has been reported that enables the simultaneous complexation of both Fe(II) and Fe(III) in the sample followed by retention of the complexes on selective solid sorbents w18x. Fe(II) was complexed by addition of 1,10-phenanthroline, while Fe(III) formed a complex with ferron (8-hydroxy-7-iodoquinoline-5-sulfonic acid). The solution was thus passed successively to an anion-exchange resin and a reversed-phase sorbent. Since the Fe(III)-ferron complex is negatively charged, it was retained by the first solid phase (for sample pH 3–6), while the Fe(II)-phenanthroline was passed through due to its non-polar character. This complex was then retained by the second solid sorbent (C18-silica). This method enabled the determination of both labile Fe(II) and Fe(III) species in wine samples. V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 4.4. Lead Lead is a toxic metal, which accumulates in the vital organs of man and animals. Its cumulative poisoning effects are serious haematological damage, anemia, kidney malfunctioning, brain damage, etc. Lead is still emitted into the biosphere in considerable amounts owing to its application as a fuel additive, mainly as tetraethyllead and tetramethyllead. It is also present in many industrial streams. Due to the presence of lead in environmental samples at low levels, its separation from other elements present and the use of a preconcentration step prior to lead determination are usually necessary. All the studies reported until now focused on the preconcentration of inorganic lead Pb(II). The performances of Pb determination using different sorbent materials have been reviewed recently w50,176x. Some examples are given below. Firstly, as a cationic species, Pb(II) can be retained on cation-exchangers, such as basic alumina w239x. However, such sorbents are rather non-selective. So, other strategies have been used that are the retention of Pb complexes on hydrophobic sorbents, the retention of Pb(II) on sorbents coated with a chelating reagent and the synthesis of new chelating resins. Lead complexed with DDTP could be retained on C18-silica, activated carbon and PUF w63x. In that way analysis through a FI system coupled to F-AAS was performed with limits of detection of 0.3, 1.2 and 3 mgyl for C18-silica, PUF and activated carbon, respectively. In another FI system Pb(II) was complexed with APDC retained on PTFE turnings and further eluted with IBMK leading to 0.8 mgyl as a limit of detection w50x. This method enabled the determination of Pb(II) in various environmental and biological samples. Complexes can be formed with DDTC in acidic medium, further retained on C18-silica and eluted with IBMK or methanol before F-AAS, enabling the detection of 3 to 10 mgyl w62,233x. Alternatively, the sorbent can be impregnated with the chelating reagent. Hence, the SPE of lead on C18silica disks has been reported after impregnation of the sorbent with a S-containing Schiff’s base to enable Pb(II) chelation w120x. After elution with nitric acid lead was further analyzed by F-AAS 1219 with a limit of detection of 16.7 myl. Retention of lead was quite selective even in the presence of other ions. The addition to the sample of ammonia as a masking agent was recommended to suppress the interfering effect of some cations, namely Cu(II), Zn(II) and Hg(II). Similarly, the coating of C18-silica with BAS was found highly selective in retaining Pb(II) with subsequent elution using acetic acid and F-AAS analysis leading to 50 ngy l as a limit of detection w121x. Pb(II) could also be retained on Amberlite XAD-7 coated with DMBS w196x or XO w176x, Amberlite IRA-904 impregnated with TCPP w8x, or PUF coated with 2-(2-benzothiazolylazo)-2-p-cresol w240x. Chelating resins have also been synthetized for the selective preconcentration of Pb(II). For example, Amberlite XAD-2 was functionalized through an azo spacer by several chelating agents, such as CA, PC and TSA, and used for the retention of Pb(II) from water samples w176x. Similarly, this sorbent functionalized with SA also enabled the retention of Pb(II), along with Zn(II) w171x. In each case nitric acid enabled lead desorption. Limits of detection in the 2.4–4.9 mgyl range were obtained. Amberlite XAD-7 functionalized with a crown ether was also found suitable for the preconcentration of Pb(II) w241x. Very recently, an attractive test procedure has been reported that enables the determination of as low as 20 ngyl of Pb(II) w203x. It is based on the immobilization of an enzyme, alkaline phosphatase on PUF. 4.5. Mercury The presence of mercury species in aquatic food chain is of great concern as it is well-known that inorganic mercury (Hg2q) is converted into highly toxic methylmercury (MeHgq) by aquatic organisms. Due to the presence of mercury in environmental samples at low levels, its separation from other elements present and the use of a preconcentration step prior to the determination is usually necessary. 4.5.1. Hg(II) Mercury can be preconcentrated from aqueous samples using a chelating ion-exchange resin con- 1220 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 taining histidine covalently bound to its carboxyl group w242x. A chelating PS-DVB-based resin with picolinic acid amide (PAA) as the functional group was also found efficient for Hg(II) retention from water samples w184x. Several functional groups chemically bound to silica gel have also been reported to afford selective sorbents for preconcentration of Hg(II) during adsorption. This was the case with DZ w104x and dithioacetal derivatives w105x. Another procedure involving the use of C18-silica disks impregnated with hexathia-18crown-6-tetraone (HT18C6TO), was shown to quantitatively extract Hg(II) from natural waters in less than 15 min w130x. Recovery was nearly independent of pH (in the range of 1 to 7) as already reported for the solvent extraction of metals with crown ethers. Before eluting Hg2q with 1 M HBr, a washing step with 1 M HNO3 was recommended to remove small amounts of retained Cu2q, Zn2q, Pb2q and Cd2q. 4.5.2. Hg(II)yorganic Hg A chelating resin based on vinyl polymer and containing dithiocarbamate groups was found efficient for retention of inorganic and organic mercury from water samples over a broad range of pHs (1 to 11) w189x. The species were eluted using an acidic aqueous solution of 5% (wyv) thiourea, enabling a preconcentration factor of 666 and the determination of mercuric species with a limit of detection of 0.2 ngyl. The use of another chelating resin containing dithiocarbamate groups was also found effective in retaining inorganic as well as organic mercury in the pH range 1–4 resulting in limit of detection of 0.5 ngyl w243x. The on-line FI preconcentration of mercury species (Hg(II), methylmercury, ethylmercury) on a dithiocarbamate resin has also been reported w244x. Detection limits of 0.05 ngyl and 0.15 ngyl for organic and inorganic mercury, respectively, could be obtained. However, this method involved manual steps once the species were eluted with thiourea, extraction into toluene as the diethyl dithiocarbamate complexes, butylation with a Grignard reagent and subsequent gas chromatography (GC) analysis. In addition, preconcentration failed in the presence of high amounts of humic substances in the water samples. Complexation of mercuric species with APDC, FI on-line preconcentration on C18-silica, further separation with LC, and analysis by CV-AAS has been reported w43x. In this manner, Hg species (methylmercury, ethylmercury, phenylmercury and Hg(II)) in fish and human urine could be analyzed at the ngyl level. Chelation with DDTP and preconcentration on C18-silica in a FI–CV-AAS system resulted in detection limits of 10 ngyl for methylmercury and Hg(II) with an enrichment factor of 20 w64x. Several ligands were tested: DDTC, APDC and DZ (or diphenylthiocarbamate) w55x. Results showed the superiority of carbamate type reagents for the preconcentration of Hg(II) and methylmercury using this system. With DDTC, detection limits of 16 ngyl of Hg could be obtained. 4.6. Selenium Selenium is present in the environment from both natural and anthropogenic inputs. This element has been recognized as an essential nutrient. However, at concentrations higher than 130 mgyl it becomes toxic. Se(IV) and Se(VI) are the predominant species in natural waters. Biomethylation may also occur, leading to the formation of organic species such as dimethyl selenide (DMSe), dimethyl diselenide (DMDSe) and diethyl selenide (DESe), and detoxification. Therefore, a reliable speciation procedure is required to evaluate toxicity of samples. An overview of SPE procedures developed for selenium has been recently given w193x and some are detailed below. It appears that most methods were dedicated to the determination of inorganic selenium. 4.6.1. Se(IV) APDC has been loaded on C18-silica and used for the preconcentration of Se from sea water prior to ET-AAS w114x. Speciation of inorganic Se(IV) and Se(VI) was possible since APDC selectively chelates Se(IV). A reduction of Se(VI) to Se(IV) prior to chelation is required for the determination of total inorganic Se. However, this method was not selective as other trace elements were retained on the sorbent (such as Bi, Pb, Zn, As, Sn, V). Complexation of Se(IV) could also be obtained V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 with 3-phenyl-5-mercapto-1,3,4-thiadiazole-2(3H)-thione (Bismuthiol II), which was then retained on activated carbon in batch experiments w222x. Se(IV) could also be retained on Chelex100 (in the iron form) over a wide pH range (up to 10) w245x. 4.6.2. Se(IV)ySe(VI) Both forms of inorganic selenium can be retained on acidic alumina and extracted from natural waters without any pH adjustment w111x. Selective elution is achieved using ammonia at different concentrations for eluting Se(VI) and then Se(IV) enabling speciation of inorganic selenium in natural waters. Quantitative recoveries from spiked tap water and ground water were obtained except for one tap water, where 42 and 154% of Se(IV) and Se(VI), respectively, were recovered possibly due to oxidation of Se(IV) by residual chlorine. Similar observations were made in another study w40x. Dowex-lX8 ion-exchanger enabled preconcentration of both Se(IV) and Se(VI) w246x. The species were then separated during elution with two different concentrations of HCl. Inorganic selenium species were also retained on several anion-exchange resins based on either cellulose or PS-DVB copolymer w225x. Despite a lower affinity, the functionalized cellulose sorbent was preferred, as it enabled a better separation of Se(IV) and Se(VI) during elution with different concentrations of nitric acid. Traces of Se(IV) and Se(VI) could also be retained on a molybdate-form anionexchange resin in batch experiments w247x, permitting speciation of inorganic selenium in natural waters except sea water (where foreign ions interfered). The coupling of SPE to LC–ICP–MS enabled the speciation of inorganic selenium in natural water samples by ion-pairing (with tetrabutylammonium phosphate as the reagent) and sorption on C18-silica, subsequent elution and separation in the chromatographic column by anion-exchange w40x. However, sample volumes were limited to 10 ml due to the limited capacity of the preconcentration column. 1221 4.6.3. Organic Se Tanzer and Heumann w248x developed a method for the selective determination of acidicyneutral and basic organoselenium species in water samples based on the selective retention of the species on Amberlite XAD-2 at different sample pHs (3 and 8, respectively). However, this method was found questionable by other authors who noted partial retention of Se(IV) on that sorbent w225x. 4.6.4. Se(IV)ySe(VI)yorganic Se Simultaneous preconcentration of inorganic and organic selenium species is a more difficult task. A combined SPE method has been reported to enable preconcentration of both inorganic and organic species of selenium w19x. An anionexchanger cartridge was placed on the top of a C18-silica cartridge so that inorganic selenium was retained on the first cartridge, while organic species were retained on the reversed-phase sorbent. The cartridges were then separated and the species eluted: 1 M HCOOH for Se(IV), 3 M HCl for Se(VI), CS2 for organic compounds. In another recent procedure, simultaneous determination of organic (selenocystine) and inorganic selenium (Se(IV) and Se(VI)) species was achieved using off-line preconcentration on Amberlite IRA-743 followed by separation of the species and analysis using LC–ICP–MS w193x. No adjustment of the sample pH was required. However, the sorbent was not selenium selective, and selenomethionine could not be retained under the conditions developed due to strong competition with hydrogen carbonate. 4.7. Tin Organotin compounds have found widespread industrial applications as biocides, antifouling paints, catalysts and polyvinyl chloride stabilizers. So, there are a variety of pathways for their entry into the environment. Whereas inorganic tin is basically harmless, some organotin compounds are highly toxic, especially tri-substituted organotin species. Therefore, there is a need for sensitive methods that enable the determination and speci- 1222 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 ation of these compounds. Several studies report the use of SPE for organotin determination w249– 251x. 4.7.1. Butyltins Tributyltins have been used as insecticides, fungicides, acaricides and preservatives for many different types of materials. In particular, they have been used as antifouling paints (as biocides) on ships, boats and dock resulting in release of TBTs directly into the aquatic environment where they are non-specific and extremely toxic to non-target animal and plants species. C18-silica, either in cartridges or in PTFE disks, has been found effective in preconcentrating TBTs from aqueous solutions w24x. Enrichment factors up to 1000 can be obtained enabling the quantification of TBTs at the 0.1 mgyl level. In addition, TBTs can be stored on such solid supports at room temperature for at least 1 month. Amberlite XAD-2 impregnated with tropolone has also been reported to retain TBT and dibutyltin, while monobutyltin was not retained w164x. The addition of 0.8% sulfuric acid to the water sample enabled the selective retention of only TBT on the resin. This species was subsequently eluted with IBMK and analyzed by ET-AAS enabling quantitative determination of TBT in water samples with a limit of detection of 14.4 ngyl and a preconcentration factor of 80. 4.7.2. Phenyltins Triphenyltin is retained on C18-silica cartridges, even though a fraction of the compound cannot be recovered w112x. Hence, best recoveries (between 81 and 89%) were obtained by elution in the backflush mode with 10y4 M 3-hydroxyflavone in methanol. Retention of TPhT on other silica-based sorbents (octyl, phenyl and cyanopropyl) has also been reported w112x. 4.7.3. Butyltinsyphenyltins A semi-automated system was reported that enabled organotin speciation w251x, wherein preconcentration was performed using a microcolumn of C18-silica placed in a FI manifold. Percolation of a derivatising reagent (sodium tetraethylborate) through the column enabled derivatisation of the organotin compounds. Elution of the derivatised species (monobutyltin, monophenyltin, dibutyltin, diphenyltin, TBT and TPhT) was achieved using methanol and their separation and analysis was performed using GC–AES. In this manner, with rather small sample volumes (10–50 ml) detection limits in the range of 0.10–0.17 ngyl could be obtained. Application of the method was tested with real river water samples, and results were consistent with those obtained using classical liquid–liquid extraction. Similarly, in an off-line system, organotins derivatised by sodium tetraethylborate could be retained on C18-silica disks, and further eluted with supercritical CO2 w252x. The use of tropolone-loaded C18-silica has also been used for the retention of several butyltins and phenyltins (mono-, di-, tri- and some tetra-substituted compounds) w118x. However, selectivity could not be achieved using SPE only. Organotins were separated by subsequent GC analysis after their ethylation with Grignard reagent. This method affords a sensitivity of low ngyl in surface waters and mgykg in sewage sludges. Speciation of tin has also been reported using both graphitized carbon black and silica gel as sorbents w223x. Water samples were first passed through GCB, allowing the retention of TBT and TPhT, while inorganic tin passed unimpeded and was analyzed directly. The organic species were subsequently eluted with a mixture of methanoly dichloromethane (4:1), and separated on silica gel. The overall method provided an enrichment factor of up to 80 000, but required the complete evaporation of the organotin fraction eluted from GCB, and the dissolution of the solid residue in hexane before separation on silica gel. Finally, organotins can be efficiently retained on strong cation-exchange silica-based bonded phases (Bond-Elut SCX) and strong cation-exchange polymeric-based phase (Oasis-MCX). The presence of NHq was essential for elution of the 4 compounds. Recoveries were lower with the polymeric phase particularly for TPhT, probably because of strong interaction of the aromatic rings with the N-vinylpyrrolidone-divinylbenzene support w36x. V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 5. Conclusion The use of SPE procedures has been growing in the past few years due to their advantages offered for trace element determinations, namely conservation of species, good preconcentration factors (thus enabling the achievement of very low limits of detection), ease of automation, and possible online coupling to instrumental techniques. Despite the numerous steps and parameters used to enable efficient extraction and recovery of the target analytes, the choice of the solid sorbent is the most critical step. Among the numerous sorbents that have been used, it clearly appears that the initial use of ion-exchangers is being replaced by more selective supports containing chelating functional groups. Such sorbents are frequently based on hydrophobic supports namely C18-silica or PS-DVB copolymer, the latter affording a broader pH range. The simplest procedure consists of adding the chelating reagent directly to the sample. A more suitable way of proceeding is to load the reagent on the solid sorbent. The coated sorbent thus obtained may usually be used several times, but partial leaching of the chelating agent may occur over time during elution of the analytes, { especially if the eluting solvent is too concentrated. Alternatively, the reagent may also be chemically bound to the sorbent leading to the synthesis of new phases, and thus avoiding any leaching of the reagent. Promising results have also been recently obtained with inorganic oxides such as titania and alumina, as retention of some trace elements could be achieved with the raw sample (i.e. no reagent addition nor pH adjustment), thereby avoiding possible speciation changes in the sample. Examples given in this review for several trace elements show the high potential of SPE, especially its possible high selectivity (by choosing the nature of the sorbent andyor the chelating agent, as well as the nature of the eluent). In fact, in some cases, differentiation of species may be achieved, thereby offering new opportunities for speciation. There is thus no doubt that this technique will face a growing interest for trace element determination and speciation in the future, as already evidenced for organic micropollutant determinations in the recent years. 1223 References w1x I. Liska, Fifty years of solid-phase extraction in water analysis-historical development and overview, J. Chromatogr. A 885 (2000) 3. w2x C.F. Poole, Solid-phase extraction, Encyclopedia of Separation Science, 3, Academic Press, 2000, p. 1405. w3x M.C. Hennion, Sample handling strategies for the analysis of non-volatile organic compounds from environmental water samples, Trends Anal. Chem. 10 (1991) 317. w4x J.L. Lundgren, A.A. Schilt, Analytical studies and applications of ferroin type chromogens immobilized by adsorption on a styrene-divinylbenzene copolymer, Anal. Chem. 49 (1977) 974. w5x M. Chikuma, M. Nakayama, T. Itoh, H. Tanaka, K. Itoh, Chelate-forming resins prepared by modification of anion-exchange resins, Talanta 27 (1980) 807. w6x H. Akaiwa, H. Kawamoto, K. Ogura, Kinetic studies of ion-exchange of cobalt(II) and nickel(II) on a resin loaded with 5-sulfo-8-quinolinol, Talanta 28 (1981) 337. w7x K.S. Lee, W. Lee, D.W. Lee, Selective separation of metal ions by a chelating agent-loaded anion exchanger, Anal. Chem. 50 (1978) 255. w8x K. Kilian, K. Pyrzynska, Preconcentration of metal ions on porphyrin-modified sorbents as pretreatment step in AAS determination, Fresenius J. Anal. Chem. 371 (2001) 1076. w9x M.E. Mahmoud, Silica gel-immobilized Eriochrome black-T as a potential solid phase extractor for zinc(II) and magnesium(II) from calcium(II), Talanta 45 (1997) 309. w10x M. Pesavento, R. Biesuz, M. Gallorini, A. Profumo, Sorption mechanism of trace amounts of divalent metal ions on a chelating resin containing iminodiacetate groups, Anal. Chem. 65 (1993) 2522. w11x M. Pesavento, R. Biesuz, J.L. Cortina, Sorption of metal ions on a weak acid cation-exchange resin containing carboxylic groups, Anal. Chim. Acta 298 (1994) 225. w12x H. Kumagai, Y. Inoue, T. Yokoyama, T.M. Suzuki, T. Suzuki, Chromatographic selectivity of rare earth elements on iminodiacetate-type chelating resins having spacer arms of different lengths: importance of steric flexibility of functional group in a polymer chelating resin, Anal. Chem. 70 (1998) 4070. w13x J. Seneviratne, J.A. Cox, Sol–gel materials for the solid phase extraction of metals from aqueous solution, Talanta 52 (2000) 801. w14x M.C. Carson, Ion-pair solid-phase extraction, J. Chromatogr. A 885 (2000) 343. w15x V. Porta, E. Mentasti, C. Sarzanini, M.C. Gennaro, Ionpair liquid–solid extraction for the preconcentration of trace metal ions, Talanta 35 (1988) 167. w16x P. Janos, K. Stulik, V. Pacakova, An ion-exchange separation of metal cations on a C-18 column coated with dodecylsulfate, Talanta 39 (1992) 29. 1224 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 w17x R. Van Grieken, Preconcentration methods for the analysis of water by X-ray spectrometric techniques, Anal. Chim. Acta 143 (1982) 3. w18x A.B. Tawali, G. Schwedt, Combination of solid phase extraction and flame atomic absorption spectrometry for differentiated analysis of labile iron(II) and iron(III) species, Fresenius J. Anal. Chem. 357 (1997) 50. w19x J.L. Gomez-Ariza, J.A. Pozas, I. Giraldez, E. Morales, Use of solid phase extraction for speciation of selenium compounds in aqueous environmental samples, Analyst 124 (1999) 75. w20x E.M. Thurman, K. Snavely, Advances in solid-phase extraction disks for environmental chemistry, Trends Anal. Chem. 19 (2000) 18. w21x C.F. Poole, Solid-phase extraction with disks, Encyclopedia of Separation Science, 9, Academic Press, 2000, p. 4141. w22x M. Kumar, D.P.S. Rathore, A.K. Singh, Quinalizarin anchored on Amberlite XAD-2. A new matrix for solidphase extraction of metal ions for flame atomic absorption spectrometric determination, Fresenius J. Anal. Chem. 370 (2001) 377. w23x G.A. Junk, M.J. Avery, J.J. Richard, Interferences in solid-phase extraction using C-18 bonded porous silica cartridges, Anal. Chem. 60 (1988) 1347. w24x O. Evans, B.J. Jacobs, A.L. Cohen, Liquid–solid extraction of tributyltin from marine sediments, Analyst 116 (1991) 15. w25x D.F. Hagen, C.G. Markell, G.A. Schmitt, D.D. Blevins, Membrane approach to solid-phase extractions, Anal. Chim. Acta 236 (1990) 157. w26x D.T. Rossi, N. Zhang, Automating solid-phase extraction: current aspects and future prospects, J. Chromatogr. A 885 (2000) 97. w27x L.B. Bjorklund, ¨ G.M. Morrison, Determination of copper speciation in freshwater samples through SPEspectrophotometry, Anal. Chim. Acta 343 (1997) 259. w28x R.E. Majors, Sample preparation for HPLC and gas chromatography using solid-phase extraction, LG-GC 4 (1989) 972. w29x R.E. Sturgeon, S.S. Berman, A. Desaulniers, D.S. Russell, Pre-concentration of trace metals from sea-water for determination by graphite-fumace atomic-absorption spectrometry, Talanta 27 (1980) 85. w30x T.M. Florence, G.E. Batley, Trace metals species in seawater-I. Removal of trace metals from sea-water by a chelating resin, Talanta 23 (1976) 179. w31x D.M. Sanchez, R. Martin, R. Morante, J. Marin, M.L. Munuera, Preconcentration speciation method for mercury compounds in water samples using solid phase extraction followed by reversed phase high performance liquid chromatography, Talanta 52 (2000) 671. w32x M.C. Yebra, N. Carro, M.F. Enriquez, A. Moreno-Cid, A. Garcia, Field sample pre-concentration of copper in sea water using chelating minicolumns subsequently incorporated on a flow-injection-flame atomic absorption spectrometry system, Analyst 126 (2001) 933. w33x E. Castillo, J.-L. Cortina, J.-L. Beltran, M.-D. Prat, M. Granados, Simultaneous determination of Cd(II), Cu(II) and Pb(II) in surface waters by solid phase extraction and flow injection analysis with spectrophotometric detection, Analyst 126 (2001) 1149. w34x M. Shamsipur, A.R. Ghiasvand, H. Sharghi, H. Naeimi, Solid phase extraction of ultra trace copper(II) using octadecyl silica membrane disks modified by a naphthol-derivative Schiff’s base, Anal. Chim. Acta 408 (2000) 271. w35x M.E. Mahmoud, E.M. Soliman, Silica-immobilized formylsalicylic acid as a selective phase for the extraction of iron(III), Talanta 44 (1997) 15. w36x E. Gonzalez-Toledo, M. Benzi, R. Compano, M. Granados, M.D. Prat, Speciation of organotin compounds in shellfish by liquid chromatography–fluorimetric detection, Anal. Chim. Acta 443 (2001) 183. w37x O. Keil, J. Dahmen, D.A. Volmer, Automated matrix separation and preconcentration for the trace level determination of metal impurities in ultrapure inorganic salts by high-resolution ICP–MS, Fresenius J. Anal. Chem. 364 (1999) 694. w38x J.F. Tyson, Flow injection atomic spectrometry, Spectrochim. Acta Rev. 14 (1991) 169. w39x M. Bittner, J.A.C. Broekaert, Speciation of chromium by solid-phase extraction coupled to reversed-phase liquid chromatography with UV detection, Anal. Chim. Acta 364 (1998) 31. w40x Y. Cai, M. Cabanas, J.L. Fenandez-Turiel, M. Abalos, J.M. Bayona, On-line preconcentration of selenium(IV) and selenium(VI) in aqueous matrices followed by liquid chromatography-inductively coupled plasma mass spectrometry determination, Anal. Chim. Acta 314 (1995) 183. w41x Q. Hu, G. Yang, J. Yin, Y. Yao, Determination of trace lead, cadmium and mercury by on-line column enrichment followed by RP-HPLC as metal-tetra-(4-bromophenyl)-porphyrin chelates, Talanta 57 (2002) 751. w42x B.W. Wenclawiak, T. Hees, C.E. Zoller, ¨ H.-P. Kabus, Rhodium and palladium b-diketonate determination with on-line supercritical fluid extraction-high performance liquid chromatography via solid phase extraction, Fresenius J. Anal. Chem. 358 (1997) 471. w43x X. Yin, W. Frech, E. Hoffmann, C. Ludke, ¨ J. Skole, Mercury speciation by coupling cold vapour atomic absorption spectrometry with flow injection on-line preconcentration and liquid chromatographic separation, Fresenius J. Anal. Chem. 361 (1998) 761. w44x S. Olsen, L.C.R. Pessenda, J. Ruzicka, E.H. Hansen, Combination of flow injection analysis with flame atomic absorption spectrophotometry: determination of trace amounts of heavy metals in polluted seawater, Analyst 108 (1983) 905. w45x Z. Fang, J. Ruzicka, E.H. Hansen, An efficient flowinjection system with on-line ion-exchange preconcentration for the determination of trace amounts of heavy V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 w46x w47x w48x w49x w50x w51x w52x w53x w54x w55x w56x w57x w58x metals by atomic absorption spectrometry, Anal. Chim. Acta 164 (1984) 23. Z. Fang, Z. Zhu, S. Zhang, S. Xu, L. Guo, L. Sun, Online separation and preconcentration in flow injection analysis, Anal. Chim. Acta 214 (1988) 41. Z. Fang, M. Sperling, B. Welz, Flow injection on-line sorbent extraction preconcentration for graphite furnace atomic absorption spectrometry, J. Anal. At. Spectrom. 5 (1990) 639. A.N. Anthemidis, G.A. Zachariadis, J.A. Stratis, Online solid phase extraction system using PTFE packed column for the flame atomic absorption spectrometric determination of copper in water samples, Talanta 54 (2001) 935. A.N. Anthemidis, G.A. Zachariadis, J.-S. Kougoulis, J.A. Stratis, Flame atomic absorption spectrometric determination of chromium(VI) by on-line preconcentration system using a PTFE packed column, Talanta 57 (2002) 15. G.A. Zachariadis, A.N. Anthemidis, P.G. Bettas, J.A. Stratis, Determination of lead by on-line solid phase extraction using a PTFE micro-column and flame atomic absorption spectrometry, Talanta 57 (2002) 919. M. Zougagh, A. Garcia de Torres, J.M. Cano Pavon, Determination of cadmium in water by ICP–AES with on-line adsorption preconcentration using DPTH-gel and TS-gel microcolumns, Talanta 56 (2002) 753. D.A. Weeks, K.W. Bruland, Improved method for shipboard determination of iron in seawater by flow injection analysis, Anal. Chim. Acta 453 (2002) 21. Q. Pu, Q. Sun, Z. Hu, S. Su, Application of 2mercaptobenzothiazole-modified silica gel to on-line preconcentration and separation of silver for its atomic absorption spectrometric determination, Analyst 123 (1998) 239. S. Zhang, Q. Pu, P. Liu, Q. Sun, Z. Su, Synthesis of amidinothioueido–silica gel and its application to flame atomic absorption spectrometric determination of silver, gold and palladium with on-line preconcentration and separation, Anal. Chim. Acta 452 (2002) 223. M. Fernandez Garcia, R. Pereiro Garcia, N. Bordel Garcia, A. Sanz-Medel, On-line preconcentration of inorganic mercury and methylmercury in sea-water by sorbent-extraction and total mercury determination by cold vapour atomic absorption spectrometry, Talanta 41 (1994) 1833. Z.-S. Liu, S.-D. Huang, Automatic on-line preconcentration system for graphite furnace atomic absorption spectrometry for the determination of trace metals in sea water, Anal. Chim. Acta 281 (1993) 185. A. Ali, X. Yin, H. Shen, Y. Ye, X. Gu, 1,10-Phenanthroline as a complexing agent for on-line sorbent extractionypreconcentration for flow injection-flame atomic absorption spectrometry, Anal. Chim. Acta 392 (1999) 283. P.-G. Su, S.-D. Huang, Use of 4-(2-pyridylazo)resorcinol or 2-(2-pyridylazo)-5-dimethylamino- w59x w60x w61x w62x w63x w64x w65x w66x w67x w68x w69x w70x 1225 phenol as chelating agent for determination of cadmium in seawater by atomic absorption spectrometry with online flow-injection sorbent extraction, Anal. Chim. Acta 376 (1998) 305. T. Prasada Rao, S. Karthikeyan, B. Vijayalekshmy, C.S.P. Iyer, Speciative determination of chromium(VI) and chromium(III) using flow-injection on-line preconcentration and flame atomic-absorption spectrometric detection, Anal. Chim. Acta 369 (1998) 69. M. Sperling, X. Yin, B. Welz, Differential determination of chromium(VI) and total chromium in natural waters using flow injection on-line separation and preconcentration electrothermal atomic absorption spectrometry, Analyst 117 (1992) 629. Y. Ye, A. Ali, X. Yin, Cobalt determination with FIFAAS after on-line sorbent preconcentration using lnitroso-2-naphthol, Talanta 57 (2002) 945. R. Lima, K.C. Leandro, R.E. Santelli, Lead preconcentration onto C18-minicolumn in continuous flow and its determination in biological and vegetable samples by flame atomic absorption spectrometry, Talanta 43 (1996) 977. S.P. Quinaia, J.B.B. Da Silva, M.C.E. Rollemberg, A.J. Curtius, Preconcentration of lead complexed with O,Odiethyl-dithiophosphate by column solid-phase extraction using different sorbents in a flow injection system coupled to a flame atomic absorption spectrometer, Talanta 54 (2001) 687. A.C.P. Monteiro, L.S.N. Andrade, R.C. Campos, Online mercury and methylmercury pre-concentration by adsorption of their dithiophosphoric acid diacylester chelates on a C18 column and cold-vapor atomic-absorption detection, Fresenius J. Anal. Chem. 371 (2001) 353. J. Ruzicka, A. Arndal, Sorbent extraction in flow injection analysis and its application to enhancement of atomic spectrometry, Anal. Chim. Acta 216 (1989) 243. Z.F. Queiroz, F.R.P. Rocha, G. Knapp, F.J. Krug, Flow system with in-line separationypreconcentration coupled to graphite furnace atomic absorption spectrometry with W–Rh permanent modifier for copper determination in seawater, Anal. Chim. Acta 463 (2002) 275. L.S.G. Teixeira, F.R.P. Rocha, M. Korn, B.F. Reis, S.L.C. Ferreira, A.C.S. Costa, Flow-injection solid-phase spectrophotometry for the determination of zinc in pharmaceutical preparations, Anal. Chim. Acta 383 (1999) 309. ´ S. Blain, P. Treguer, Iron(II) and iron(III) determination in sea water at the nanomolar level with selective online preconcentration and spectrophotometric determination, Anal. Chim. Acta 308 (1995) 425. E. Vassileva, N. Furuta, Application of high-surfacearea ZrO2 in preconcentration and determination of 18 elements by on-line flow injection with inductively coupled plasma atomic emission spectrometry, Fresenius J. Anal. Chem. 370 (2001) 52. A.G. Cox, I.G. Cook, C.W. McLeod, Rapid sequential determination of chromium(III)–chromium(VI) by flow 1226 w71x w72x w73x w74x w75x w76x w77x w78x w79x w80x w81x V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 injection analysis-inductively coupled plasma atomicemission spectrometry, Analyst 110 (1985) 331. M. Sperling, S. Xu, B. Welz, Determination of chromium(III) and chromium(VI) in water using flow injection on-line preconcentration with selective adsorption on activated alumina and flame atomic absorption spectrometric detection, Anal. Chem. 64 (1992) 3101. M.J. Marques, A. Morales-Rubio, A. Salvador, M. de la Guardia, Chromium speciation using activated alumina microcolumns and sequential injection analysisflame atomic absorption spectrometry, Talanta 53 (2001) 1229. A.G. Cox, C.W. McLeod, Preconcentration and determination of trace chromium(III) by flow injectiony inductively-coupled plasmayatomic emission spectrometry, Anal. Chim. Acta 179 (1986) 487. S. Nielsen, E.H. Hansen, Selective flow-injection quantification of ultra-trace amounts of Cr(VI) via on-line complexation and preconcentration with APDC followed by determination by electrothermal atomic absorption spectrometry, Anal. Chim. Acta 366 (1998) 163. W. Som-Aum, S. Liawruangrath, E.H. Hansen, Flow injection on-line preconcentration of low levels of Cr(VI) with detection by ETAAS. Comparison of using an open tubular PTFE knotted reactor and a column reactor packed with PTFE beads, Anal. Chim. Acta 463 (2002) 99. C. Shuyu, Z. Zhifeng, Y. Huaming, Dithione as chelator in the flow injection separation and pre-concentration system of trace metals in biological samples, Anal. Chim. Acta 451 (2002) 305. Z. Fang, S. Xu, L. Dong, W. Li, Determination of cadmium in biological materials by flame atomic absorption spectrometry with flow-injection on-line sorption preconcentration, Talanta 41 (1994) 2165. R. Compano, R. Ferrer, J. Guiteras, M.D. Prat, Spectrofluorimetric detection of zinc and cadmium with 8(benzene sulfonamido)-quinoline immobilized on a polymeric matrix, Analyst 119 (1994) 1225. I.A. Kovalev, L.V. Bogacheva, G.I. Tsysin, A.A. Formanovsky, Y.A. Zolotov, FIA-FAAS system including on-line solid phase extraction for the determination of palladium, platinum and rhodium in alloys and ores, Talanta 52 (2000) 39. P. Hashemi, A. Olin, Equilibrium and kinetic properties of a fast iminodiacetate based chelating ion exchanger and its incorporation in a FIA-ICP–AES system, Talanta 44 (1997) 1037. R.M. Cespon-Romero, M.C. Yebra-Biurrun, M.P. Bermejo-Barrera, Preconcentration and speciation of chromium by the determination of total chromium and chromium(III) in natural waters by flame atomic absorption spectrometry with a chelating ion-exchange flow injection system, Anal. Chim. Acta 327 (1996) 37. w82x N.G. Beck, R.P. Francks, K.W. Bruland, Analysis of Cd, Cu, Ni, Zn and Mn in estuarine water by inductively coupled plasma mass spectrometry coupled with an automated flow injection system, Anal. Chim. Acta 455 (2002) 11. w83x S. Matsuoka, Y. Tennichi, K. Takehara, K. Yoshimura, Flow analysis of micro amounts of chromium(III) and (VI) in natural water by solid phase spectrophotometry using diphenylcarbazide, Analyst 124 (1999) 787. w84x C.E.S. Miranda, B.F. Reis, N. Baccan, A.P. Packer, M.F. ´ Automated flow analysis system based on multiGine, commutation for Cd, Ni and Pb on-line pre-concentration in a cationic exchange resin with determination by inductively coupled plasma atomic emission spectrometry, Anal. Chim. Acta 453 (2002) 301. w85x K. Yoshimura, S. Matsuoka, Y. Inokura, U. Hase, Flow analysis for trace amounts of copper by ion-exchanger phase absorptiometry with 4,7-diphenyl-2,9-dimethyl1,10-phenanthroline disulfonate and its application to the study of karst groundwater storm runoff, Anal. Chim. Acta 268 (1992) 225. w86x J.A. Gomes Neto, A.P. Oliveira, G.P.G. Feshi, C.S. Dakuzaku, M. Moraes, Minimization of lead and copper interferences on spectrophotometric determination of cadmium using electrolytic deposition and ion-exchange in multi-commutation flow system, Talanta 53 (2000) 497. w87x J.B.B. Silva, M.B.O. Giacomelli, A.J. Curtius, Determination of bismuth in aluminium and in steels by electrothermal atomic absorption spectrometry after online separation using a minicolumn of activated carbon, Analyst 124 (1999) 1249. w88x R.E. Santelli, M. Gallego, M. Valcarcel, Preconcentration and atomic absorption determination of copper traces in waters by on-line adsorption-elution on an activated carbon microcolumn, Talanta 41 (1994) 817. w89x H. Zhang, X. Yuan, X. Zhao, Q. Jin, On-line preconcentration with activated carbon for microwave plasma torch atomic emission spectrometry, Talanta 44 (1997) 1615. w90x S. Lin, C. Zheng, G. Yun, Determination of palladium by flame atomic absorption spectrometry combined online with flow injection preconcentration using a microcolumn packed with activated carbon fibre, Talanta 42 (1995) 921. w91x S.K. Xu, M. Sperling, B. Welz, Flame atomic absorption spectrometric determination of cadmium and copper in biological reference materials using on-line sorbent extraction preconcentration, Fresenius J. Anal. Chem. 344 (1992) 535. w92x C.S.L. Ferreira, D.S. Jesus, R.J. Cassella, A.C.S. Costa, M.S. Carvalho, R.E. Santelli, An on-line solid phase extraction system using polyurethane foam for the spectrophotometric determination of nickel in silicates and alloys, Anal. Chim. Acta 378 (1999) 287. w93x S. Tsakovski, K. Benkhedda, E. Ivanova, F.C. Adams, Comparative study of 8-hydroxyquinoline derivatives as V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 w94x w95x w96x w97x w98x w99x w100x w101x w102x w103x w104x w105x w106x w107x chelating reagents for flow-injection preconcentration of cobalt in a knotted reactor, Anal. Chim. Acta 453 (2002) 143. S.D. Hartenstein, J. Ruzicka, G.D. Christian, Sensitivity enhancement for flow injection analysis-inductively coupled plasma atomic emission spectrometry using an on-line preconcentration ion-exchange column, Anal. Chem. 57 (1985) 21. K. Benkhedda, H.G. Infante, F.C. Adams, E. Ivanova, Inductively coupled plasma mass spectrometry for trace analysis using flow injection on-line preconcentration and time-of-flight mass analyser, Trends Anal. Chem. 21 (2002) 332. J.N. King, J.S. Fritz, Concentration of metal ions by complexation with sodium bis(2-hydroxyethyl)dithiocarbamate and sorption on XAD-4 resin, Anal. Chem. 57 (1985) 1016. C. Kantipuly, S. Katragadda, A. Chow, H.D. Gesser, Chelating polymers and related supports for separation and preconcentration of trace metals, Talanta 37 (1990) 491. C. Gueguen, J. Dominck, D. Perret, Use of chelating resins and inductively coupled plasma mass spectrometry for simultaneous determination of trace and major elements in small volumes of saline water samples, Fresenius J. Anal. Chem. 370 (2001) 909. O. Abollino, M. Aceto, M.C. Bruzzoniti, E. Mentasti, C. Sarzanini, Speciation of copper and manganese in milk by solid-phase extractionyinductively coupled plasma-atomic emission spectrometry, Anal. Chim. Acta 375 (1998) 299. A. Tong, Y. Akama, S. Tanak, Pre-concentration of copper, cobalt and nickel with 3-methyl-l-phenyl-4stearoyl-5-pyrazolone loaded on silica gel, Analyst 115 (1990) 947. K. Terada, K. Matsumoto, Y. Taniguchi, Preconcentration of palladium(II) from water with thionalide loaded onto silica gel, Anal. Chim. Acta 147 (1983) 411. K. Terada, K. Matsumoto, T. Inaba, Differential preconcentration of arsenic(III) and arsenic(V) with thionalide loaded on silica gel, Anal. Chim. Acta 158 (1984) 207. K. Terada, K. Nakamura, Preconcentration of cobalt(II) in natural waters with 1-nitroso-2-naphthol supported on silica gel, Talanta 28 (1981) 123. M.E. Mahmoud, M.M. Osman, M.E. Amer, Selective pre-concentration and solid phase extraction of mercury(II) from natural water by silica gel-loaded dithizone phases, Anal. Chim. Acta 415 (2000) 33. M.E. Mahmoud, G.A. Gohar, Silica gel-immobilizeddithioacetal derivatives as potential solid phase extractors for mercury(II), Talanta 51 (2000) 77. M.E. Mahmoud, Selective solid phase extraction of mercury(II) by silica gel-immobilized-dithiocarbamate derivatives, Anal. Chim. Acta 398 (1999) 297. Y. Sohrin, S.-I. Iwamoto, S. Akiyama, T. Fujita, T. Kugii, H. Obata, E. Nakayama, S. Goda, Y. Fujishima, H. Hasegawa, K. Ueda, M. Matsui, Determination of w108x w109x w110x w111x w112x w113x w114x w115x w116x w117x w118x w119x w120x 1227 trace elements in seawater by fluorinated metal alkoxide glass-immobilized 8-hydroxyquinoline concentration and high-resolution inductively coupled plasma mass spectrometry detection, Anal. Chim. Acta 363 (1998) 11. E. Vassileva, K. Hadjiinov, T. Stoychev, C. Daiev, Chromium speciation analysis by solid-phase extraction on a high surface area TiO2, Analyst 125 (2000) 693. E. Vassileva, I. Proinova, K. Hadjiivanov, Solid-phase extraction of heavy metal ions on a high surface area titanium dioxide, Analyst 121 (1996) 607. A.C. Sahayam, Speciation of Cr(III) and Cr(VI) in potable waters by using activated neutral alumina as collector and ET-AAS for determination, Anal. Bioanal. Chem. 372 (2002) 840. K. Pyrzynska, P. Drzewicz, M. Trojanowicz, Preconcentration and separation of inorganic selenium species on activated alumina, Anal. Chim. Acta 363 (1998) 141. R. Compano, M. Granados, C. Leal, M.D. Prat, Solidphase extraction and spectrofluorimetric determination of triphenyltin in environmental samples, Anal. Chim. Acta 283 (1993) 272. J. Saurina, C. Leal, R. Compano, M. Granados, R. Tauler, M.D. Prat, Determination of triphenyltin in seawater by excitation–emission matrix fluorescence and multivariate curve resolution, Anal. Chim. Acta 409 (2000) 237. R.E. Sturgeon, S.N. Willie, S.S. Berman, Preconcentration of selenium and antimony from seawater for determination by graphite furnace atomic absorption spectrometry, Anal. Chem. 57 (1985) 6. R.E. Sturgeon, S.S. Berman, S.N. Willie, Concentration of trace metals from sea-water by complexation with 8hydroxyquinoline and adsorption on C18-bonded silica gel, Talanta 29 (1982) 167. Y. Yamini, A. Tamaddon, Solid-phase extraction and spectrophotometric determination of trace amounts of copper in water samples, Talanta 49 (1999) 119. Y. Yamini, N. Amiri, Solid-phase extraction, separation, and visible spectrophotometric determination of trace amounts of iron in water samples, J. AOAC Int. 84 (2001) 713. ¨ M.D. Muller, Comprehensive trace level determination of organotin compounds in environmental samples using high-resolution gas chromatography with flame photometric detection, Anal. Chem. 59 (1987) 617. M. Shamsipur, A. Avanes, M.K. Rofouei, H. Sharghi, G. Aghapour, Solid-phase extraction and determination of ultra trace amounts of copper(II) using octadecyl silica membrane disks modified by 11-hydroxynaphthacene-5,12-quinone and flame atomic absorption spectrometry, Talanta 54 (2001) 863. O.R. Hashemi, M.R. Kargar, F. Raoufi, A. Moghimi, H. Aghabozorg, M.R. Ganjali, Separation and preconcentration of trace amounts of lead on octadecyl silica membrane disks modified with a new S-containing 1228 w121x w122x w123x w124x w125x w126x w127x w128x w129x w130x w131x V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 Schiff’s base and its determination by flame atomic absorption spectrometry, Microchem. J. 69 (2001) 1. M. Shamsipur, F. Raoufi, H. Sharghi, Solid phase extraction and determination of lead in soil and water samples using octadecyl silica membrane disks modified by biswl-hydroxy-9,10-anthraquinone-2-methylxsulphide and flame atomic absorption spectrometry, Talanta 52 (2000) 637. D.W. King, J. Lin, D.R. Kester, Spectrophotometric determination of iron(II) in seawater at nanomolar concentrations, Anal. Chim. Acta 247 (1991) 125. Z. Yi, G. Zhuang, P.R. Brown, R.A. Duce, Highperformance liquid chromatographic method for the determination of ultratrace amounts of iron(II) in aerosols, rainwater, and seawater, Anal. Chem. 64 (1992) 2826. Z.-S. Liu, S.-D. Huang, Determination of copper and cadmium in sea water by preconcentration and electrothermal atomic absorption spectrometry, Anal. Chim. Acta 267 (1992) 31. M. Shamsipur, A.R. Ghiasvand, Y. Yamini, Solid-phase extraction of ultratrace uranium(VI) in natural waters using octadecyl silica membrane disks modified by trin-octylphosphine oxide and its spectrophotometric determination with dibenzoylmethane, Anal. Chem. 71 (1999) 4892. M.B. Shabani, T. Akagi, A. Masuda, Preconcentration of trace rare-earth elements in seawater by complexation with bis(2-ethylhexyl) hydrogen phosphate and 2-ethylhexyl dihydrogen phosphate adsorbed on a C18 cartridge and determination by inductively coupled plasma mass spectrometry, Anal. Chem. 64 (1992) 737. Y. Yamini, M. Chaloosi, H. Ebrahimzadeh, Solid phase extraction and graphite furnace atomic absorption spectrometric determination of ultra trace amounts of bismuth in water samples, Talanta 56 (2002) 797. Y. Yamini, J. Hassan, R. Mohandesi, N. Bahramifar, Preconcentration of trace amounts of beryllium in water sample on octadecyl silica cartridges modified by quinalizarine and its determination with atomic absorption spectrometry, Talanta 56 (2002) 375. M. Shamsipur, M.H. Mashhadizadeh, Preconcentration of trace amounts of silver ion in aqueous samples on octadecyl silica membrane disks modified with hexathia-18-crown-6 and its determination by atomic absorption spectrometry, Fresenius J. Anal. Chem. 367 (2000) 246. Y. Yamini, N. Alizadeh, M. Shamsipur, Solid phase extraction and determination of ultra trace amounts of mercury(II) using octadecyl silica membrane disks modified by hexathia-18-crown-6-tetraone and cold vapour atomic absorption spectrometry, Anal. Chim. Acta 355 (1997) 69. R.J. Kvitek, J.F. Evans, P.W. Carr, Diamineysilanemodified controlled pore glass. The covalent attachment reaction from aqueous solution and the mechanism of w132x w133x w134x w135x w136x w137x w138x w139x w140x w141x w142x w143x w144x reaction of bound diamine with copper(II), Anal. Chim. Acta 144 (1982) 93. A.R. Sarkar, P.K. Datta, M. Sarkar, Sorption recovery of metal ions using silica gel modified with salicylaldoxime, Talanta 43 (1996) 1857. K. Terada, K. Matsumoto, H. Kimura, Sorption of copper(II) by some complexing agents loaded on various supports, Anal. Chim. Acta 153 (1983) 237. R.E. Sturgeon, S.S. Berman, S.N. Willie, J.A.H. Desaulniers, Preconcentration of trace elements from seawater with silica-immobilized 8-hydroxyquinoline, Anal. Chem. 53 (1981) 2337. J.W. McLaren, A.P. Mykytiuk, S.N. Willie, S.S. Berman, Determination of trace metals in seawater by inductively coupled plasma mass spectrometry with preconcentration on silica-immobilized 8-hydroxyquinoline, Anal. Chem. 57 (1985) 2907. R. Kocjan, S. Przeszlakowski, Calcon-modified silica gel sorbent. Application to preconcentration or elimination of trace metals, Talanta 39 (1992) 63. D.E. Leyden, G.H. Luttrell, A.E. Sloan, N.J. DeAngelis, Characterization and application of silylated substrates for the preconcentration of cations, Anal. Chim. Acta 84 (1976) 97. T. Seshadri, H.-J. Haupt, Silica-immobilized 2-w(2-(triethoxysilyl)ethyl)thioxaniline as a selective sorbent for the separation and preconcentration of palladium, Anal. Chem. 60 (1988) 47. P. Jones, P.J. Hobbs, L. Ebdon, A dithizone post-column detector for the high-performance liquid chromatographic determination of trace metals, Anal. Chim. Acta 149 (1983) 39. M.E. Mahmoud, M.S.M. Al Saadi, Selective solid phase extraction and preconcentration of iron(III) based on silica gel-chemically immobilized purpurogallin, Anal. Chim. Acta 450 (2001) 239. K.S. Abou-El-Sherbini, I.M.M. Kenawy, M.A. Hamed, R.M. Issa, R. Elmorsi, Separation and preconcentration in a batch mode of Cd(II), Cr(III, VI), Cu(II), Mn(II, VII) and Pb(II) by solid-phase extraction by using of silica modified with N-propylsalicylaldimine, Talanta 58 (2002) 289. P. Lessi, N.L. Dias Filho, J.C. Moreira, J.T.S. Campos, Sorption and preconcentration of metal ions on silica gel modified with 2,5-dimercapto-1,3,4-thiadiazole, Anal. Chim. Acta 327 (1996) 183. R. Garcia-Valls, A. Hrdlicka, J. Perutka, J. Havel, N.V. Deorkard, L.L. Tavlarides, M. Munoz, M. Valiente, Separation of rare earth elements by high performance liquid chromatography using a covalent modified silica gel column, Anal. Chim. Acta 439 (2001) 247. S. Hutchinson, G.A. Kearney, E. Horne, B. Lynch, J.D. Glennon, M.A. McKervey, S.J. Harris, Solid phase extraction of metal ions using immobilized chelating calixarene tetrahydroxamates, Anal. Chim. Acta 291 (1994) 269. V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 w145x R.M. Izatt, R.L. Bruening, M.L. Bruening, B.J. Tarbet, K.E. Krakowiak, J.S. Bradshaw, J.J. Christensen, Removal and separation of metal ions from aqueous solutions using a silica-gel-bonded macrocycle system, Anal. Chem. 60 (1988) 1825. w146x M.L. Bruening, D.M. Mitchell, J.S. Bradshaw, R.M. Izatt, R.L. Bruening, Effect of organic solvent and anion type on cation binding constants with silica gel bound macrocycles and their use in designing selective concentrator columns, Anal. Chem. 63 (1991) 21. w147x G. Philippeit, J. Angerer, Determination of palladium in human urine by high-performance liquid chromatography and ultraviolet detection after ultraviolet photolysis and selective solid-phase extraction, J. Chromatogr. B 760 (2001) 237. w148x V. Cuculic, M. Mlakar, M. Branica, Synergetic adsorption of copper(II) mixed ligand complexes onto the SEP-PAK C18 column, Anal. Chim. Acta 339 (1997) 181. w149x M.H. Pournaghi-Azar, Dj. Djozan, H.A. Zadeh, Determination of trace bismuth by solid phase extraction and anodic stripping voltammetry in non-aqueous media, Anal. Chim. Acta 437 (2001) 217. w150x S.J. Kumar, P. Ostapczuk, H. Emons, Chromium speciation in water by electrothermal AAS after simultaneous adsorption of Cr(III) and Cr(VI) on activated alumina mini column, At. Spectrosc. 20 (1999) 194. w151x J.L. Manzoori, M.H. Sorouraddin, F. Shemirani, Chromium speciation by a surfactant-coated alumina microcolumn using electrothermal atomic absorption spectrometry, Talanta 42 (1995) 1151. w152x E. Vassileva, K. Hadjiinov, Determination of trace elements in AR grade alkali salts after preconcentration by column solid-phase extraction on TiO2 (anatase), Fresenius J. Anal. Chem. 357 (1997) 881. w153x P. Liang, Y. Qin, B. Hu, C. Li, T. Peng, Z. Jiang, Study on the adsorption behavior of heavy metal ions on nanometer-size titanium dioxide with ICP–AES, Fresenius J. Anal. Chem. 368 (2000) 638. w154x C.W. Huck, G.K. Bonn, Recent developments in polymer-based sorbents for solid-phase extraction, J. Chromatogr. A 885 (2000) 51. w155x N. Masque, ´ R.M. Marce, ´ F. Borrull, New polymeric and other types of sorbents for solid-phase extraction of polar organic micropollutants from environmental water, Trends Anal. Chem. 17 (1998) 384. w156x L. Elci, L.M. Soylak, A. Uzun, E. Buyukpatir, ¨ ¨ M. Dogan, Determination of trace impurities in some nickel compounds by flame atomic absorption spectrometry after solid phase extraction using Amberlite XAD-16 resin, Fresenius J. Anal. Chem. 368 (2000) 358. w157x O. Abollino, M. Aceto, M.C. Bruzzoniti, E. Mentasti, C. Sarzanini, Determination of metals in highly saline matrices by solid-phase extraction and slurry-sampling inductively coupled plasma–atomic emission spectrometry, Anal. Chim. Acta 375 (1998) 293. 1229 w158x S. Saracoglu, ¸ ¸ Column solid-phase extraction L. Elci, with Chromosorb-102 resin and determination of trace elements in water and sediment samples by flame atomic absorption spectrometry, Anal. Chim. Acta 452 (2002) 77. w159x K. Isshiki, Y. Sohrin, H. Karatani, E. Nakayama, Preconcentration of chromium(III) and chromium(VI) in sea water by complexation with quinolin-8-ol and adsorption on macroporous resin, Anal. Chim. Acta 224 (1989) 55. w160x O. Abollino, E. Mentasti, V. Porta, C. Sarzanini, Immobilized 8-oxine units on different solid sorbents for the uptake of metal traces, Anal. Chem. 62 (1990) 21. w161x A. Tunceli, ¨ ¸ A.R. Turker, Speciation of Cr(III) and Cr(VI) in water after preconcentration of its 1,5-diphenylcarbazone complex on amberlite XAD-16 resin and determination by FAAS, Talanta 57 (2002) 1199. w162x A.G. Howard, M.H. Arbab-Zavar, The preconcentration of mercury and methylmercury on dithizone-coated polystyrene beads, Talanta 26 (1979) 895. w163x P. Bermejo-Barrera, G. Gonzalez-Campos, M. FerronNovais, A. Bermejo-Barrera, Column preconcentration of organotin with tropolone-immobilized and their determination by electrothermal atomization absorption spectrometry, Talanta 46 (1998) 1479. w164x P. Bermejo-Barrera, R.M. Anllo-Sendin, M.J. CantelarBarbazan, A. Bermejo-Barrera, Selective preconcentration and determination of tributyltin in fresh water by electrothermal atomic absorption spectrometry, Anal. Bioanal. Chem. 372 (2002) 837. w165x C.S.L. Ferreira, C.F. de Brito, A.F. Dantas, N.M. Lopo de Araujo, A.C.S. Costa, Nickel determination in saline matrices by ICP–AES after sorption on Amberlite XAD-2 loaded with PAN, Talanta 48 (1999) 1173. w166x K. Isshiki, F. Tsuji, T. Kuwamoto, E. Nakayama, Preconcentration of trace metals from seawater with 7dodecenyl-8-quinolinol impregnated macroporous resin, Anal. Chem. 59 (1987) 2491. w167x A. Ramesh, K.R. Mohan, K. Seshaiah, Preconcentration of trace metals on Amberlite XAD-4 resin coated with dithiocarbamates and determination by inductively coupled plasma–atomic emission spectrometry in saline matrices, Talanta 57 (2002) 243. w168x A.N. Masi, R.A. Olsina, Preparation and characterization of chelating resins loaded with 2-(5-bromo-2pyridylazo)-5-(diethylamino)phenol for preconcentration of rare earth elements, Fresenius J. Anal. Chem. 357 (1997) 65. w169x M.E. Leon-Gonzalez, L.V. Perez-Arribas, Chemically modified polymeric sorbents for sample preconcentration, J. Chromatogr. A 902 (2000) 3. w170x R. Saxena, A.K. Singh, S.S. Sambi, Synthesis of a chelating polymer matrix by immobilizing Alizarin RedS on Amberlite XAD-2 and its application to the preconcentration of lead(II), cadmium(II), zinc(II) and nickel(II), Anal. Chim. Acta 295 (1994) 199. 1230 V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 w171x R. Saxena, A.K. Singh, D.P.S. Rathore, Salicylic acid functionalized polystyrene sorbent Amberlite XAD-2. Synthesis and applications as a preconcentrator in the determination of zinc(II) and lead(II) by using atomic absorption spectrometry, Analyst 120 (1995) 403. w172x P.K. Tewari, A.K. Singh, Thiosalicylic acid-immobilized Amberlite XAD-2: metal sorption behavior and applications in estimation of metal ions by flame atomic absorption spectrometry, Analyst 125 (2000) 2350. w173x R. Saxena, A.K. Singh, Pyrocatechol violet immobilized Amberlite XAD-2: synthesis and metal-ion uptake properties suitable for analytical applications, Anal. Chim. Acta 340 (1997) 285. w174x P.K. Tewari, A.K. Singh, Amberlite XAD-2 functionalized with chromotropic acid: synthesis of a new polymer matrix and its applications in metal ion enrichment for their determination by flame atomic absorption spectrometry, Analyst 124 (1999) 1847. w175x P.K. Tewari, A.K. Singh, Synthesis, characterization and applications of pyrocatechol modified amberlite XAD2 resin for preconcentration and determination of metal ions in water samples by flame atomic absorption spectrometry, Talanta 53 (2001) 823. w176x P.K. Tewari, A.K. Singh, Preconcentration of lead with Amberlite XAD-2 and Amberlite XAD-7 based chelating resins for its determination by flame atomic absorption spectrometry, Talanta 56 (2002) 735. w177x M. Kumar, D.P.S. Rathore, A.K. Singh, Metal ion enrichment with Amberlite XAD-2 functionalized with Tiron: analytical applications, Analyst 125 (2000) 1221. w178x K. Dev, G.N. Rao, Preparation and analytical properties of a chelating resin containing bicine groups, Talanta 42 (1995) 591. w179x M.C. Yebra-Biurrun, M.C. Garcia-Dopazo, A. BermejoBarrera, M.P. Bermejo-Barrera, Preconcentration of trace amounts of manganese from natural waters by means of a macroreticular poly(dithiocarbamate)resin, Talanta 39 (1992) 671. w180x N. Uehara, A. Katamine, Y. Shijo, High-performance liquid chromatographic determination of cobalt(II) as the 2-(5-bromo-2-pyridylazo)-5-diethylaminophenol chelate after preconcentration with a cation-exchange resin, Analyst 119 (1994) 1333. w181x B.C. Mondal, D. Das, A.K. Das, Synthesis and characterization of a new resin functionalized with 2-naphthol3,6-disulfonic acid and its application for the speciation of chromium in natural water, Talanta 56 (2002) 145. w182x Y. Cai, G. Jiang, J. Liu, Preconcentration of cobalt with 8-hydroxyquinoline and gas chromatographic stationary phase Chromosorb 105 and its determination by graphite furnace atomic absorption spectrometry, Talanta 57 (2002) 1173. w183x F. Shemirani, M. Rajabi, Preconcentration of chromium(III) and speciation of chromium by electrothermal atomic absorption spectrometry using cellulose adsorbent, Fresenius J. Anal. Chem. 371 (2001) 1037. w184x B. Sengupta, J. Das, Preconcentration of trace amounts of mercury(II) in water on picolinic acid amide-containing resin, Anal. Chim. Acta 219 (1989) 339. w185x P.K. Tewari, A.K. Singh, Amberlite XAD-7 impregnated with Xylenol Orange; a chelating collector for preconcentration of Cd(II), Co(II), Cu(II), Ni(II), Zn(II) and Fe(III) ions prior to their determination by flame AAS, Fresenius J. Anal. Chem. 367 (2000) 562. w186x B. Wen, X.-Q. Shuan, J. Lian, Separation of Cr(III) and Cr(VI) in river and reservoir water with 8-hydroxyquinoline immobilized polyacrylonitrile fiber for determination by inductively coupled plasma mass spectrometry, Talanta 56 (2002) 681. w187x B. Gong, Y. Wang, ICP–AES determination of traces of noble metal ions pre-concentrated and separated on a new polyacrylacylaminothiourea chelating fiber, Anal. Bioanal. Chem. 372 (2002) 597. w188x B. Wen, X.-Q. Shuan, R.-X. Liu, H.-X. Tang, Preconcentration of trace elements in sea water with poly(acrylaminophosphonic-dithiocarbamate) chelating fiber for their determination by inductively coupled plasma mass spectrometry, Fresenius J. Anal. Chem. 363 (1999) 251. w189x K. Minagawa, Y. Takizawa, I. Kifune, Determination of very low levels of inorganic and organic mercury in natural waters by cold-vapor atomic absorption spectrometry after preconcentration on a chelating resin, Anal. Chim. Acta 115 (1980) 103. w190x S. Blain, P. Appriou, H. Handel, Preconcentration of trace metals from sea water with the chelating resin Chelamine, Anal. Chim. Acta 272 (1993) 91. w191x S.-C. Pai, P.-Y. Whung, R.-L. Lai, Pre-concentration effiency of Chelex-100 resin for heavy metals in seawater, Anal. Chim. Acta 211 (1988) 257. w192x M.E. Malla, M.B. Alvarez, D.A. Batistoni, Evaluation of sorption and desorption characteristics of cadmium, lead and zinc on Amberlite IRC-718 iminodiacetate chelating ion exchanger, Talanta 57 (2002) 277. w193x M. Bueno, M. Potin-Gautier, Solid-phase extraction for the simultaneous preconcentration of organic (selenocystine) and inorganic wSe(IV), Se(VI)x selenium in natural waters, J. Chromatogr. A 963 (2002) 185. w194x M. Llobat-Estelles, A.R. Mauri-Aucejo, M.D. LopezCatalan, Spectrophotometric determination of chromium with diphenylcarbazide in the presence of vanadium, molybdenum, and iron after separation by solid-phase extraction, Fresenius J. Anal. Chem. 371 (2001) 358. w195x A. Afkhami, T. Madrakian, A.A. Assl, { A.A. Sehhat, Solid phase extraction flame atomic absorption spectrometric determination of ultra-trace beryllium, Anal. Chim. Acta 437 (2001) 17. w196x S. Hoshi, H. Fujisawa, K. Nakamura, S. Nakata, M. Uto, K. Akatsuka, Preparation of Amberlite XAD resins coated with dithiosemicarbazone compounds and preconcentration of some metal ions, Talanta 41 (1994) 503. V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 w197x B. Salih, R. Say, A. Denizli, O. Gene, E. Piskin, Determination of inorganic and organic mercury compounds by capillary gas chromatography coupled with atomic absorption spectrometry after preconcentration on dithizone-anchored poly(ethylene glycol dimethacrylate-hydroxyehtylmethacrylate) microbeads, Anal. Chim. Acta 371 (1998) 177. w198x D.W. Lee, M. Halmann, Selective separation of nickel(II) by dimehtylglyoxime-treated polyurethane foam, Anal. Chem. 48 (1976) 2214. w199x T. Braun, A.B. Farag, Plasticized open-cell polyurethane foam as a universal matrix for organic reagents in trace element preconcentration. Part III. Collection of cobalt traces on 1-nitroso-2-naphthol and diethyldithiocarbamate foams, Anal. Chim. Acta 76 (1975) 107. w200x A. Alexandrova, S. Arpadjan, Determination of trace elements in analytical-reagent grade sodium salts by atomic absorption spectrometry and inductively coupled plasma atomic emission spectrometry after preconcentration by column solid phase extraction, Analyst 118 (1993) 1309. w201x D.S. Jesus, R.J. Cassella, S.L.C. Ferreira, A.C.S. Costa, M.S. Carvalho, R.E. Santelli, Polyurethane foam as a sorbent for continuous flow analysis: preconcentration and spectrophotometric determination of zinc in biological materials, Anal. Chim. Acta 366 (1998) 263. w202x C.S.L. Ferreira, H.C. Santos, D.S. Jesus, Molybdenum determination in iron matrices by ICP–AES after separation and preconcentration using polyurethane foam, Fresenius J. Anal. Chem. 369 (2001) 187. w203x I.A. Veselova, T.N. Shekhovtsova, Visual determination of lead(II) by inhibition of alkaline phosphatase immobilized on polyurethane foam, Anal. Chim. Acta 413 (2000) 95. w204x S. Tian, G. Schwedt, Solid-phase extraction of the chromium(III)-diphenylcarbazone complex prior to ionpair chromatography and application to geological samples, Fresenius J. Anal. Chem. 354 (1996) 447. w205x I.E. De Vito, R.A. Olsina, A.N. Masi, Enrichment method for trace amounts of rare earth elements using chemofiltration and XRF determination, Fresenius J. Anal. Chem. 368 (2000) 392. w206x J.P. Riley, D. Taylor, Chelating resins for the concentration of trace elements from sea water and thier analytical use in conjunction with atomic absorption spectrophotometry, Anal. Chim. Acta 40 (1968) 479. w207x R. Boniforti, R. Ferraroli, P. Frigieri, D. Heltai, G. Queirazza, Intercomparison of five methods for the determination of trace metals in sea water, Anal. Chim. Acta 162 (1984) 33. w208x M. Pesavento, R. Biesuz, Sorption of divalent metal ions on an iminodiacetic resin from artificial seawater, Anal. Chim. Acta 346 (1997) 381. w209x S.-C. Pai, Pre-concentration efficiency of Chelex-100 resin for heavy metals in seawater. Part 2. Distribution of heavy metals on a Chelex-100 column and optimi- w210x w211x w212x w213x w214x w215x w216x w217x w218x w219x w220x w221x w222x 1231 zation of the column efficiency by a plate simulation method, Anal. Chim. Acta 211 (1988) 271. C. Gueguen, C. Belin, B.A. Thomas, F. Monna, P.-Y. Favarger, J. Dominik, The effect of freshwater UVirradiation prior to resin preconcentration of trace metals, Anal. Chim. Acta 386 (1999) 155. H. Kumagai, T. Yokoyama, T.M. Suzuki, T. Suzuki, Liquid chromatographic selectivity and retention behavior of rare-earth elements on a chelating resin having a propylenediaminetetraacetate type functional group, Analyst 124 (1999) 1595. B. Wen, X.-Q. Shuan, S.-G. Xu, Preconcentration of ultratrace rare earth elements in seawater with 8-hydroxyquinoline immobilized polyacrylonitrile hollow fiber membrane for determination by inductively coupled plasma mass spectrometry, Analyst 124 (1999) 621. M.R. Buchmeiser, R. Tessadri, G. Seeber, G.K. Bonn, Selective extraction of rare-earth elements from rocks using a high-capacity cis-1,4-butanedioic acid-functionalized resin, Anal. Chem. 70 (1998) 2130. F. Sinner, M.R. Buchmeiser, R. Tessadri, M. Mupa, K. Wurst, G.K. Bonn, Dipyridyl amide-functionalized polymers prepared by ring-opening-metathesis polymerization (ROMP) for the selective extraction of mercury and palladium, J. Am. Chem. Soc. 120 (1998) 2790. D.J. Hutchinson, A.A. Schilt, Investigation of the adsorption of ferroin-type ligands and metal chelates on activated carbons for applications in reagent purification and trace metal enrichment and determination, Anal. Chim. Acta 154 (1983) 159. A. Afkhami, T. Madrakian, Kinetic-spectrophotometric determination of selenium in natural water after preconcentration of elemental selenium on activated carbon, Talanta 58 (2002) 311. E. Piperaki, H. Berndt, E. Jackwerth, Investigations on the sorption of metal chelates on activated carbon, Anal. Chim. Acta 100 (1978) 589. Y. Petit de Pena, M. Gallego, M. Varcarcel, Flame atomic absorption spectrometric determination of cadmium in biological samples using a preconcentration flow system with an activated carbon column and dithizone as a chelating agent, J. Anal. At. Spectrom. 9 (1994) 691. T. Aydemir, S. Gucer, Determination of nickel in urine by flame atomic absorption spectrometry after activated carbon enrichment, Anal. Lett. 29 (1996) 351. M. Soylak, I. Narin, L. Elci, M. Dogan, Atomic absorption-spectrometric determination of copper, cadmium, lead and nickel in urine samples after enrichment and separation procedure on an activated carbon column, Trace Elem. Electrolytes 16 (1999) 131. ¨ M. Yaman, S. Gucer, Determination of cadmium and lead in vegetables after activated-carbon enrichment by atomic absorption spectrometry, Analyst 120 (1995) 101. T. Kubota, K. Suzuki, T. Okutani, Determination of total selenium content in sediments and natural water 1232 w223x w224x w225x w226x w227x w228x w229x w230x w231x w232x w233x w234x w235x w236x V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 by graphite furnace-atomic absorption spectroscopy after collection as a selenium(IV) complex on activated carbon, Talanta 42 (1995) 949. T. Ferri, E. Cardarelli, B.M. Petronio, Determination of tin and triorganotin compounds in sea-water by graphitefurnace atomic-absorption spectrophotometry, Talanta 36 (1989) 513. M. Soylak, U. Divrikli, L. Elci, M. Dogan, Preconcentration of Cr(III), Co(II), Cu(II), Fe(III) and Pb(II) as calmagite chelates on cellulose nitrate membrane filter prior to their flame atomic absorption spectrometric determinations, Talanta 56 (2002) 565. K. Pyrzynska, Separation of inorganic selenium species on anion-exchange resins, Analyst 120 (1995) 1933. A. Bhalotra, B.K. Puri, Preconcentration of bismuth(III) and copper(II) by solid-phase extraction and subsequent determination by differential pulse polarography, J. AOAC Intern. 84 (2001) 47. B.K. Puri, M. Satake, G. Kano, S. Usami, Selective preconcentration of iron in beverages and water samples using 2,4,6-tri-2-pyridinyl-1,3,5-triazine-tetraphenylborate-naphthalene adsorbent, Anal. Chem. 59 (1987) 1850. B.K. Puri, S. Balani, Preconcentration of iron(III), cobalt(II) and copper(II) nitroso-R-complexes on tetradecyldimehtylbenzylammonium iodie-naphthalene adsorbent, Talanta 42 (1995) 337. C.F. Poole, A.D. Gunatilleka, R. Sethuraman, Contributions of theory to method development in solid-phase extraction, J. Chromatogr. A 885 (2000) 17. M. Pesavento, E. Baldini, Study of sorption of copper(II) on complexing resin columns by solid phase extraction, Anal. Chim. Acta 389 (1999) 59. K. Hirose, Y. Dokiya, Y. Sugimura, Determination of conditional stability constants of organic copper and zinc complexes dissolved in seawater using ligand exchange method with EDTA, Mar. Chem. 11 (1982) 343. P.J.M. Buckley, C.M.G. van den Berg, Copper complexation profiles in the Atlantic ocean. A comparative study using electrochemical and ion exchange techniques, Mar. Chem. 19 (1986) 281. R. Ma, W. Van Mol, F. Adams, Determination of cadmium, copper and lead in environmental samples. An evaluation of flow injection on-line sorbent extraction for flame atomic absorption spectrometry, Anal. Chim. Acta 285 (1994) 33. W. Frenzel, Highly sensitive semi-quantitative field test for the determination of chromium(VI) in aqueous samples, Fresenius J. Anal. Chem. 361 (1998) 774. S. Ahmad, R.C. Murthy, S.V. Chandra, Chromium speciation by column chromatography using a direct plasma atomic emission spectrometer, Analyst 115 (1990) 287. I. Kubrakova, T. Kudinova, A. Formanovsky, N. Kuz’min, G. Tsysin, Y. Zolotov, Determination of chromium(III) and chromium(VI) in river water by electrothermal atomic absorption spectrometry after w237x w238x w239x w240x w241x w242x w243x w244x w245x w246x w247x w248x w249x sorption preconcentration in a microwave field, Analyst 119 (1994) 2477. D.M. Adria-Cerezo, M. Llobat-Estelles, A.R. MauriAucejo, Preconcentration and speciation of chromium in waters using solid-phase extraction and atomic absorption spectrometry, Talanta 51 (2000) 531. M.M. Gibbs, A simple method for the rapid determination of iron in natural waters, Water Res. 13 (1979) 295. Y. Zhang, P. Riby, A.G. Cox, C.W. McLeod, A.R. Date, Y.Y. Cheung, On-line pre-concentration and determination of lead in potable water by flow injection atomic absorption spectrometry, Analyst 113 (1988) 125. V.A. Lemos, S.L.C. Ferreira, On-line preconcentration system for lead determination in seafood samples by flame atomic absorption spectrometry using polyurethane foam loaded with 2-(2-benzothiazolylazo)-2-pcresol, Anal. Chim. Acta 441 (2001) 281. P. Sooksamiti, H. Geckeis, K. Grudpan, Determination of lead in soil samples by in-valve solid-phase extraction-flow injection flame atomic absorption spectrometry, Analyst 121 (1996) 1413. C.-Y. Liu, Histidine as the functional group for a chelating ion exchanger, Anal. Chim. Acta 192 (1987) 85. E. Yamagami, S. Tateishi, A. Hashimoto, Application of a chelating resin to the determination of trace amounts of mercury in natural waters, Analyst 105 (1980) 491. H. Emteborg, D.C. Baxter, W. Frech, Speciation of mercury in natural waters by capillary gas chromatography with a microwave-induced plasma emission detector following preconcentration using a dithiocarbamate resin microcolumn installed in a closed flow injection system, Analyst 118 (1993) 1007. T. Ferri, P. Sangiorgio, Determination of selenium speciation in river waters by adsorption on iron(III)Chelex-100 resin and differential pulse cathodic stripping voltammetry, Anal. Chim. Acta 321 (1996) 185. ¨ U. Ornemark, A. Olin, Preconcentration and separation of inorganic selenium on Dowex IX8 prior to hydride generation-atomic absorption spectrometry, Talanta 41 (1994) 67. T. Kubota, T. Okutani, Determination of selenium content in natural water by grphite furnace atomic absorption spectrometry after preconcentration with molybdate-form anion exchange resin, Anal. Chim. Acta 351 (1997) 319. D. Tanzer, K.G. Heumann, Determination of dissolved selenium species in environmental water samples using isotope dilution mass spectrometry, Anal. Chem. 63 (1991) 1984. W.M.R. Dirkx, R. Lobinski, F.C. Adams, Speciation analysis of organotin in water and sediments by gas chromatography with optical spectrometric detection V. Camel / Spectrochimica Acta Part B 58 (2003) 1177–1233 after extraction separation, Anal. Chim. Acta 286 (1994) 309. w250x J.L. Gomez-Ariza, E. Morales, I. Giraldez, D. SanchezRodas, A. Velasco, Sample treatment in chromatography-based speciation of organometallic pollutants, J. Chromatogr. A 938 (2001) 211. w251x J. Szpunar-Lobinska, M. Ceulemans, R. Lobisnki, F.C. Adams, Flow-injection sample preparation for organotin 1233 speciation analysis of water by capillary gas chromatography-microwave-induced plasma atomic emission spectrometry, Anal. Chim. Acta 278 (1993) 99. w252x Y. Cai, J.M. Bayona, Simultaneous speciation of butyl-, phenyl-, and cyclohexyltin compounds in aqueous matrices using ethylation followed by solid-phase trace enrichment, SFE, and GC determination, J. Chromatogr. Sci. 33 (1995) 89.