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 Cadmium 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 Cadmium. Finally, examples of applications are presented. From the analytical tools above listed, FAAS presents low costs, operational facility and high sample throughput. The determination of cadmium by flame atomic spectrometry is free of interference and this can be easily atomized in air-acetylene flame. In the resonance line 228 nm, the characteristic concentration is 0.02 mg L−1. The analytical line at 326.1 nm is suitable for determining higher Cd concentrations and the characteristic concentration is about 6mg L−1, so that excessive dilution can be avoided [15]. Cadmium can be determined by CV AAS where atomic Cd vapor is measured by AAS in an unheated quartz cell. LOD of 80 ng L−1 was obtained which can be further improved by working at low temperature. The sensitivity can be increased collecting the atomic Cd vapor in a graphite tube pre-treated with palladium at 150 ◦C and then reatomizing at 1600 ◦C [14]. Cold vapor generation coupled to atomic absorption spectrometry with flow injection (FI-CV AAS) was evaluated as a rapid and simple method for the determination of cadmium [16]. The determination of cadmium by ETAAS was, for a long period, difficulty because the cadmium is an element with high volatility [17].

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 [1,2,3]. However, numerous studies have also shown the great potential of this technique for speciation studies [18,161-177].

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 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.

1.2 Retention of trace Cadmium on the sorbent

Adsorption of trace Cadmium 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 Cadmium, ion-pair solid phase extraction may be used.

1.2.1 Adsorption.

Trace Cadmium 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.

1.2.2 Chelation.

Several functional group atoms are capable of chelating trace Cadmium. 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 trace Cadmium. 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 (Ca²⁺, Mg²⁺, Na⁺) that form rather weak outer-sphere complexes with only hard oxygen ligands.

– group II-‘borderline’ cations: these have an intermediate character; this group contains Fe , Co²⁺  Ni²⁺  Cu²⁺  Zn²⁺  Pb²⁺  Mn²⁺  . They possess affinity for both hard and soft ligands. – group III-‘soft’ cations: these tend to form covalent bonds. Hence, Cd²⁺ and Hg²⁺ possess strong affinity for intermediate (N) and soft (S) ligands.

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 Cadmium, 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 simplest 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. 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) [5–8]. 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 [9]. However, the presence of the solid sorbent may also have an effect and Cadmium to the formation of a complex with a different stoichiometry than the one observed in a homogeneous reaction [10,11]. In fact, several characteristics of the sorbent should be taken into account, namely the number of active groups available in the resin phase [7,10], the length of the spacer arm between the resin and the bound ligand [12], and the pore dimensions of the resin [13].

 

2 Cadmium determination

2.1 Atomic absorption spectrometry (AAS)

The determination of Cadmium by flame atomic absorption spectrometry (FAAS) is practically free of interference and requires an air–acetylene flame [13]. The interference caused by aluminum and iron can be overcome by the addition of ascorbic acid, citric acid and EDTA [13]. The threshold of sensitivity of this technique is very low (LOD 0.01 mg L−1) and is often unsuitable for trace analysis [13]. In this sense, many preconcentration procedures must be performed to determine trace amounts of Cadmium, as indicated in the tables shown here. The most important analytical lines of Cadmium are 217.00 and 283.31 nm. The 217.00 nm line is more sensitive, notwithstanding the greater amount of background absorption effects [13].

Electrothermal atomic absorption spectrometry (ETAAS) is a good alternative for determining trace amounts of Cadmium in several types of samples in view of its good sensitivity [13]. However,

in some cases, previous preconcentration and separation steps are carried out before analytical measurements by ETAAS.

The use of a modifier stabilizes Cadmium, allowing for its determination without causing matrix effects. The Pd–Mg modifier is the one most commonly used, since it produces the best results. This modifier allows for the application of pyrolysis temperatures ranging from 1200 to 1400 ◦C, which enables the separation of most interfering elements [13–17]. The stabilizing effect of this modifier also raises the atomization temperature to 2000 ◦C, which allows a characteristic mass of about 16 pg [13]. Ammonium phosphate is another modifier frequently used for determining Cadmium by ETAAS, allowing for an atomization temperature of 1600 ◦C and enabling a low characteristic mass of 12 pg [13,18].

2.2 Inductively coupled plasma optical emission spectrometry (ICP OES)

Inductively coupled plasma optical emission spectrometry is an analytical technique often employed to determine Cadmium in various types of samples [19]. The main emission lines are: Cd II 220.353 nm, Cd I 216.999 nm and Cd I 283.306 nm, with 220.353 nm being the most sensitive. However, the low level of Cadmium in many samples lies below the detection limit of this technique. Moreover, several types of spectral interference have been reported in the determination of Cadmium by ICP OES. Virtually all photomultiplier-based ICP spectrometers use the Cd 220.353 nm analytical line, despite its severe background continuum and inter-element interference from Al 220.4 nm and background shift due to iron (Fe). Direct spectral overlap interference due to iron has also been found in the 216.9 and 283.9 nm analytical lines [20,21]. Thus, preconcentration and separation procedures have been devised to allow trace amounts of Cadmium to be determined in complex matrices using ICP OES. Several tables shown in this paper summarize the use of separation techniques such as liquid–liquid extraction, solid phase extraction, cloud point extraction and others, as pre-steps in determining Cadmium using ICP OES.

3 Cadmium separation and preconcentration

Separation techniques such as coprecipitation [23,50], liquid–liquid extraction [57–60], solid phase extraction [69–134] and more recently, cloud point extraction [139–146,150] and on-line coprecipitation using a knotted reactor [150–160] have been successfully employed to determine

trace levels of Cadmium. Each technique has its pros and cons and should be chosen according to the analytical problem.

3.1 Coprecipitation

Coprecipitation is one of the most efficient separation/ enrichment techniques for trace heavy metal ions. The main requirement for this technique is that the collector should separate easily from the matrix solution. This can be done by filtering, centrifuging and washing of the precipitate. In addition, it is desirable that the collector should be a pure and readily available substance. The advantages of this technique are its simplicity and the fact that various analyte ions can be preconcentrated and separated simultaneously from the matrix. Inorganic or organic

coprecipitants have been used as efficient collectors of trace elements. However, this process is slow and samples sometimes have to be kept over-night for complete coprecipitation [22]. This technique has been widely applied in preconcentration procedures for determining Cadmium inwater samples. Inorganic coprecipitants such as manganese dioxide and aluminum, gallium, cerium(IV), erbium, iron(III), magnesium, samarium and zirconium hydroxides have been widely and successfully used for preconcentrating trace Cadmium ions from different mediums [23,50]. A fast procedure for separation and preconcentration using ultrasound-assisted coprecipitation with manganese dioxide has recently been proposed for determining Cadmium in environmental samples by ICP OES [18]. Coprecipitation parameters, including concentration of oxidizing agent—KMnO4, concentration of MnSO4 and exposure time to ultrasound irradiation,

are discussed. The time required for coprecipitation is about 60 s.

Another paper proposed an aluminum hydroxide coprecipitation method for the determination of trace amounts of Cu, Cd and Cd by FAAS in seawater and mineral water after a 125-fold preconcentration [19]. Table 1 lists several methods for determining Cadmium in many samples, using coprecipitation as the separation and preconcentration technique

3.2 Liquid–liquid extraction

Solvent extraction has been one of the most extensively studied and widely applied methods in preconcentration and separation procedures for the determination of trace elements due to its simplicity, convenience, wide scope, etc. In this technique, the metal is distributed between two immiscible liquid phases (usually an aqueous and an organic phase). Metal ion stripping from aqueous solution to organic phase takes place after a complexation reaction. For the analytical measurement, the extracted metal ion can be directly measured in organic extract or a back extraction step is carried out in an aqueous medium, usually acid [51–53]. Separation and preconcentration procedures using solvent extraction generally result in a high enrichment factor due to the difference between the volumes of aqueous and organic phases. Although this procedure is operated in batch mode, it is time-consuming and produces large amounts of potentially toxic organic solvents as waste.  The implementation of methods in continuous mode overcomes these drawbacks. Such unit operations are, instead, carried out in flow injection (FI) and/or sequential injection (SI) systems, which, besides reducing sample and reagent consumption, allow all manipulations to be done automatically in an enclosed environment, thus minimizing the risk of sample contamination. Analytical procedures for Cadmium separation and preconcentration by solvent extraction and determination by atomic spectrometric techniques have been widely applied since these techniques were invented [54].

Procedures in flow injection systems for separation and preconcentration with the application of solvent extraction to determine metals, including Cadmium, are extensively discussed in the literature [55,56]. Table 2 lists several analytical systems proposed for Cadmium separation and preconcentration by solvent extraction and Cadmium determination using atomic spectrometric techniques.

3.3. Solid phase extraction

Solid phase extraction is based on the partition between a liquid (sample matrix) and a solid phase (sorbent). Several sorbents coupled to detection systems have been used for Cadmium preconcentration and determination. The basic approach is the contact of a liquid sample through a column, a flask, a cartridge, a tube or a disk containing an adsorbent that retains Cadmium ions. After this first step, the retained Cadmium is recovered upon elution with an appropriate solvent [68].

Sorbents used in preconcentration systems for Cadmium determination can be unloaded, loaded or chemically modified with the help of complexing reagents. Unloaded supports are potential collectors of analytes in the form of a single ion or associated with other species such as complexes. The sorbents in this class include activated carbon [69], natural adsorbents [70], Amberlite XAD resins [71–75], polyethylene [76] and others [77]. Cadmium can also be complexed with ligands loaded in several supports, such as polyurethane foam [78–80], activated carbon [81] and polymeric materials [82–84], which make for efficient preconcentration procedures. Many reagents have been used to load these supports and to retain Cadmium ions by complexation. These include: 2-(2_-thiazolylazo)-pcresol (TAC) [85], 2-propylpiperidine-1-carbodithioate [86], 2-(2-benzothiazolylazo)-2-p-cresol (BTAC) [80,87], pyrogallol red [81], 1-(2-pyridylazo)-2-naphthol (PAN) [88], dithizone [89,90] and 2-(5-bromo-2-pyridylazo)-5-diethyl-aminophenol (5-Br-PADAP) [91]. Supports can also be functionalized with chelating reagents, which render them powerful absorbents for Cadmium preconcentration. Supports such as Amberlite XAD series [92,93], silica gel [94–96] and cellulose [97,84] have been modified with reagents by several routes. Several methods have been proposed for preconcentrating Cadmium using the solid phase batch extraction procedure (Table 3).

3.4. Automation and possible on-line coupling toanalysis techniques

SPE can be easily automated, and several commercially available systems have been recently reviewed [123]. Home-made systems have also been reported [124]. 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.

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 or inductively coupled plasma mass spectrometer (ICP– MS), 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 [125].

3.4.2. On-line coupling to atomic absorption

spectrometry. Olsen et al [126]. and Fang et al[127,128] were the first to describe an on-line flow injection (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 Cadmium as a model trace element[129]. Since then, numerous papers reported FI with on-line preconcentration followed by AAS, as exemplified by determination of Cd [129]. Selected applications are reported in Table 4. The sorbent should provide for rapid sorption and desorption of the analytes to be used in FI systems [130]. 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 be used as eluting solvents Cadmiuming 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) [131,132]. 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 [135s], 0,0-diethyl-dithiophosphate (DDTP) [134]. The micro column can be inserted into the tip of the PTFE capillary in the autosampler arm of a graphite furnace atomic absorption spectrometer [131]. 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 4, such as functionalized activated carbon [134], polyurethane foam (PUF) [134], or PTFE turnings [133]. A particular knotted reactor (KR) has been recently developed, which consists of a long tube properly knotted usually made of PTFE.

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 [130]. Since then, several studies have used this coupling with different sorbents such as ZrO2 or functionalized silica gel for example [136].

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 [123]. Its coupling to FI analysis is well suited for monitoring purposes and a few studies present such systems as indicated in Table 4 [123]. 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. 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.

3.5. Cloud point extraction (CPE)

The cloud point phenomenon occurs when a nonionic or amphoteric surfactant above its critical micellar concentration (CMC) causes the separation of the original solution into two phases when heated at a characteristic temperature called cloud point temperature. Above the cloud point, micelles formed from surfactant molecules act as an organic solvent in liquid–liquid extraction and the analytes are partitioned between the micellar and aqueous phases [135–137]. Thus, metallic elements can be extracted to a surfactant-rich phase, trapped in the hydrophobic

micellar core, in the form of hydrophobic complexes that are formed between the metal ion and an appropriate chelating agent under adequate conditions. An evaluation of the partition

coefficients of ligands and complexes involved is, therefore, Surfactant solutions provide a medium that modifies the reaction ratio, equilibrium position and spectral and analytical parameters. However, the micellar phase obtained after a CPE has physical and chemical characteristics that must be taken into account in the development of an analytical procedure. Borges et al. [140], for example, determined Cadmium by preconcentration in a micellar phase of Triton X-114 and determination by ETAAS using Ir and Ru as permanent modifiers. This enabled them to reach a higher pyrolysis temperature and thus, eliminate the surfactant matrix before the atomization step without risk of analyte loss, as well as to avoid high background absorption. Table 5 shows some recent CPE applications for Cadmium determination and some analytical characteristics.

4. Conclusions

Considering the poisonous nature of Cadmium and the low concentration of this element in samples, preconcentration procedures have been devised involving separation techniques such as liquid–liquid extraction, solid phase extraction, coprecipitation and cloud point extraction. Nevertheless, each technique has its pros and cons and should therefore be chosen according to the analytical problem. Most of the proposed methods were established using solid phase extraction. In recent years, on-line systems have been preferred.

Table 3 Off-line procedures for Cadmium preconcentration using solid phase extraction.

TAN: 1-(2-Thiazolylazo)-2-naphthol; Br-PADAP: 2-(5-bromo-2-pyridylazo)-5-(diethylamino)-phenol; DDTP: O,O-diethyldithiophosphate

ng L−1.