摘要:

The ETV as a thermochemical reactor for ICP-MS sample introduction†‡ R. E. Sturgeon* and J. W. Lam Institute for National Measurement Standards, National Research Council of Canada, Ottawa, Ontario, Canada, K1A 0R9. E-mail: ralph.sturgeon@nrc.ca Received 3rd December 1998, Accepted 19th January 1999 Electrothermal vaporization (ETV) for sample introduction into (inductively coupled) plasmas has been explored for more than two decades, first for use with optical spectroscopy and subsequently with mass spectrometry.It is with the latter that its full potential has been appreciated vis-a`-vis solution sample nebulization. Tandem coupling of an ETV to a plasma source elicits a number of attractive features, not least of which is the explicit use of the device as a thermochemical reactor for in situ pretreatment of samples. This aspect of ETV use has not yet been suYciently well explored, despite an accumulated body of literature in the related field of ETAAS, where judicious selection of thermal programs and chemical modifiers has been extensively used to minimize analytical problems. Of particular interest for ETV sample introduction is the feasibility of using classical chemical modifiers or other reagents to alter the volatility of either the analyte or the concomitant matrix, thereby permitting a thermal or temporal separation of their release from the ETV surface.This approach may alleviate space charge interference eVects, minimize polyatomic ion interferences and eVectively enhance resolution, permit direct speciation of trace element fractions in samples as well as serve as a ‘crucible’ for sample preparation. The literature in this field is reviewed and examples of such applications for ICP-AES and ICP-MS detection are presented.application of heat, reagents or direct interaction with the Introduction substrate surface). It must be noted that the requirements for The term ‘electrothermal vaporizer’ can be used to encompass successful application of the ETV as an atomization source a wide range of physical platforms used for the resistive for AAS are distinctly diVerent from the objectives set for its heating of a substrate upon which a (condensed phase) analyte use as a vaporizer for sample introduction into plasma sources sample has been placed such as to result in its (rapid) release such as the inductively coupled and microwave induced into the gas phase.This process may result in the simple plasmas (ICPs and MIPs).Whereas complete atomization of vaporization of the analyte and any accompanying matrix, or the sample and confinement of the atomic vapor within the the partial or complete atomization of the analyte and/or high temperature observation volume is sought for AAS, it is matrix. Samples may be introduced in the form of solids, suYcient to ensure complete vaporization of the analyte species slurries, liquids, gases or aerosols.Although rods, boats, and its eYcient transport from the ETV to the plasma or other filaments, cups and tubes have been used as heated substrates, observation volume used for atomic spectrometry; in principle, discussion will be essentially limited to tubular graphite fur- the two-step atomizer (cup-in-tube) may be viewed as the use naces, as they are the most well-characterized and provide the of an ETV for sample introduction into an ETA.6 Formation most controllable thermal environment when heated either of gaseous analyte molecular species or the adsorption/ longitudinally or axially.1 Direct sample insertion (DSI) occlusion of molecules or atoms of analyte on/in transportable devices, while exhibiting some properties similar to ETVs, will (matrix) aerosol is the principal objective.7 The ETV may thus not be considered since, despite some potential for use of be considered one of the constituents of a tandem source8 in modifiers to alter the chemistry of the sample and micro- which the best characteristics of the device can be separately environment within the vaporizer, minimal thermal control is optimized to take advantage of its unique sample treatment available (determined by depth of insertion into the plasma2) capabilities in combination with those oVered by a variety of and application requires availability of a demountable torch.plasma sources. The atomic spectroscopy community is most familiar with An array of sample introduction techniques is currently a variant of the ETV technique pertaining to its use as an utilized for plasma spectrochemistry.9 Liquids are accommoatomizer (ETA) over the past 25 years in atomic absorption dated using pneumatic, ultrasonic, thermospray and hydraulic spectrometry (AAS)3 and applications of the ETV have been high pressure nebulizers or any of the more eYcient continuous amply demonstrated.4,5 The analysts’ freedom to select suitable microflow nebulizers (HEN, DIHEN, MCN, DIN, OCN, etc.) thermal programs in combination with the application of as well as through use of direct sample insertion devices.Solids gaseous and liquid (dissolved) reagents permits wide-ranging can be directly introduced (using DSI devices) as well as in physicochemical transformations of samples and analytes to the form of dry powders (using fluidized beds) or as laser be accomplished in situ (i.e., selective vaporization of matrices ablated or spark eroded aerosols.Gases are sampled directly and transformation of analyte species to other forms by following vapor generation (e.g., hydride formation). The ETV is unique amongst these in that this single device is able to accommodate all phases of matter: conventional dosing of †Presented at the Fifth Rio Symposium on Atomic Spectrometry, microliter volumes of liquid samples, direct weighing of pow- Cancu� n, Mexico, October 4–10, 1998. ‡© Canadian Crown copyright.ders, injection of slurries of solids10 as well as the sequestration J. Anal. At. Spectrom., 1999, 14, 785–791 785of chemically and physically generated analyte vapors and with such elements or high concentrations of matrix elements which are target analytes in subsequent samples. In general, it aerosols11–13 for subsequent introduction into plasmas. The advantages and disadvantages of the ETV as a sample is expedient to have a knowledge of the chemistries prevailing in the ETV reactor in order to elicit optimal performance introduction technique for plasma spectrochemistry are summarized in Table 1.The microsampling capability originally from this introduction technique, which likely accounts for the majority of users being those having a background in touted as a major benefit to use of the ETV has been significantly eroded in the past few years following the intro- ETAAS. Being cognizant of these limitations and shortcomings, it is duction and acceptance of numerous microflow nebulizers operating with high eYciency in the <100 ml min-1 flow useful to reflect on what the potential contributions this sample introduction technique might encompass.There is a general range.9 Nevertheless, small, discrete sample volumes can be conveniently processed with the ETV. More significantly, the reluctance to use this device for routine work and, to date, most applications have (unfortunately) focussed on liquid ETV eYciently handles solutions containing dissolved solids, up to 100% content (i.e., solids), which would plug nebulizers samples, which are often best handled by the many pneumatic microsample introduction approaches already available.As when continuously aspirated,14–20 and can be directly used for the analysis of organic solvents. The eYciency21–23 of these noted by Montaser et al.,9 ‘…the true strength of the ETV… is the ability to handle diYcult liquid matrices (organic devices is typically in the range of 20%, well above that of conventional nebulizer sample introduction, but less than that samples, high salt samples, radioactive and toxic materials), slurries, solids and to allow speciation studies’.While many available with microflow nebulizers. Enhanced detection power ensues,4,5 often as a direct result of reduced solvent-related such samples c indeed be used with the ETV, the restrictions noted above concerning formation of non-volatile species and polyatomic spectral interferences (with lower oxide fractions as well ) in addition to the possible sample preconcentration memory eVects from carbide forming elements (i.e., with radioactive elements) must be respected.ETV was first coupled schemes which can be implemented with the ETV. Furthermore, and uniquely, it is possible to enhance the to an ICP for optical emission work some 25 years ago,25 with the first report on its application to ICP-MS in 1983.26 Several eVective resolution of the MS system by taking advantage of the added temporal (thermal ) dimension associated with the reviews have already explored many niche applications of the ETV which illustrate some of its strengths;4,5,27 it is the analyte desorption event,24 utilize the device as a digestion medium for sample pretreatment and derive speciation infor- objective of this review to highlight the proficiencies of the ETV which advantageously utilize both its physical and chemi- mation.In general, sample manipulation can be minimized with this device, in that many physicochemical reactions can cal attributes to achieve unique in situ sample handling and pretreatment capabilities. Emphasis will be placed on thermo- be implemented in situ. Use of the ETV is not without significant drawbacks. The chemical interactions of the analyte–matrix with the ETV substrate and/or added gas, liquid or solid reagents. Examples discrete nature of sampling leads to transient response, which is more diYcult to quantitate precisely (5–15% RSD typical ) are drawn from relevant ETAAS literature which can, by extrapolation, be utilized for such tandem source applications, than steady-state signals, although precision can be improved to 0.5–2% by using the isotope dilution approach5.The lower as well as reports directly addressing ETV sample introduction issues. Areas relating to the use of the ETV as a preconcen- sample throughput (typically 20 per h) frequently unsettles users.The recent introduction of ICP-time of flight instrumen- tration cell (to enhance detection power further and minimize matrix interferences), for in situ sample preparation, as well tation will remove these obstacles currently associated with quadrupole based detectors, which limit acquisition to 4–5 as analyte speciation will also be addressed, although it is recognized that these topics strictly do not fall under the elements per transient and restrict measures which can be taken to enhance precision.Clearly, non-volatile elements and purview of the chemistry of the system. Issues relating to analyte transport eYciency and the need for physical carriers those that are prone to formation of low volatility carbides are not optimal candidates for quantitation by this technique. are also not considered here and the reader is referred elsewhere for a discussion of these topics.21–23 The potential for severe memory eVects remains when dealing Table 1 Advantages and disadvantages of the ETV for sample Chemistry in the ETV introduction Reduction of matrix/spectral interferences Advantages— Microsampling capability One of the key advantages of the ETV for sample introduction High dissolved solids samples accommodated into ICP-MS instruments is the use of the device to modify High sample transport eYciency the composition of the sample entering the plasma.This is Compatible with organic solvents Enhanced detection power most conveniently accomplished by application of judicious Reduced solvent spectral interferences thermal programming.Appropriate ashing temperatures Reduced analyte oxide fractions should be used to remove unwanted matrix constituents when Control of matrix (space charge) interferences possible along with the objective of minimal transport of Enhanced eVective resolution (added dimensionality) matrix components to the plasma at all times, even during the Speciation capabilities high temperature clean step.In combination with suitable In situ sample digestion Minimization of sample handling reagents to execute physicochemical reactions which eVect separation of the analyte from the matrix species, the ETV Disadvantages— functions as a dynamic, albeit, highly non-uniform, thermo- Transient signals; peak duration 2–5 s chemical reactor. It is likely that the first reference to the Limited to 4–5 elements per transient (scanning instruments) ETV as a ‘thermochemical reactor’ was coined by Gilmut- Low throughput dinov et al.28 Poor(er) precision Limited to volatile elements/species It is expedient at this point to discuss interferences and their Memory eVects origins.Clearly, if the response from an analyte is altered in Internal standard (carrier) needed a sample matrix compared with that from a standard, a matrix Method of additions often required for quantitation eVect is operative.The origin may lie within the ETV itself, Requires knowledge of chemistries in ETV reactor wherein the matrix may promote the low temperature loss of 786 J. Anal. At. Spectrom., 1999, 14, 785–791analyte (e.g., as a volatile chloride) or inhibit its complete Elimination of matrix components entering the plasma may not only alleviate spectral interferences as well as non- volatilization (by occlusion in a low volatility component). Sample transport eYciency from the ETV to the plasma may spectroscopic eVects (e.g., ionization suppression from EIEs), but the matrix eVect associated with the space charge phenom- also be altered by the presence of the matrix, serving either to enhance (through formation of a more stable transportable enon can, in principle, also be minimized.Caution must be exercised in that space charge eVects from sample components aerosol ) or decrease the eYciency (by formation of a reactive film on the surface of the transport tube).Once in the plasma, are hopefully not replaced with an equally deleterious matrix from added modifiers. An example of this will be presented matrix induced signal suppressions or enhancements may originate from an alteration of the degree of ionization, or the later in connection with the determination of Se in a reference sediment material. zone of maximum ion density sampled. Within the MS interface, space charge eVects may dominate, leading to mass As noted earlier, refractory elements, and those that interact at high temperature with graphite to form non-volatile car- bias problems, especially when dealing with relatively massive ionized matrix components and low relative atomic mass bides, are troublesome for ETV sample introduction, resulting in non-quantitative release from the ETV or excessively broad analytes.Finally, spectroscopic interferences take the form of polyatomic ions, such as matrix element argides or oxides with ‘transients’ having degraded precision of measurement. Similarly, such matrix species may be diYcult to remove from nominally the same mass as the analyte.Although the eVects of some of these may be remediated by use of sector instru- the device and create problems for subsequent determinations. This dilemma is not unique to tandem source ETV, having ments, often the maximum resolution of 10 000 available with these machines is not suYcient and semi-empirical corrections been earlier encountered with ETAAS and even dc arc emission spectrography.As a consequence, several approaches have to response can be attempted. Spectroscopic interferences may, for all practical purposes, also be isobaric (e.g., Cd and Sn been adopted which take advantage of the chemistry which can be used at elevated temperature to create conditions within isotopes) which would require resolving powers of 105 to eliminate. As most ICP-MS instruments are quadrupole based, the ETV which enhance the volatility of the sample and/or analyte, including the introduction of halogenation and com- with an eVective unit resolution, all ‘spectroscopic’ interferences become synonymous with ‘isobaric’ interferences.plexation reagents.36–47 Ng and Caruso38 reported on the use of a 7% (m/v) solution of NH4Cl to facilitate the vaporization of Zr, U, V and Cr from a graphite cup by preferential in situ Solvent removal to minimize polyatomic interferences.Simple removal of the sample solvent, typically water, can substan- formation of the volatile chlorides. Similarly, Matousek and Powell36 utilized direct injection of chlorine gas (50 ml ) into tially alter the concentrations of oxide and hydroxide species in the plasma, thereby reducing spectral interferences the furnace during the high temperature vaporization stage to achieve eYcient removal of refractory carbide residues. This from polyatomic interferences. This process is eYciently accomplished with the ETV during the sample drying stage, provided an alternative to halocarbon purging.47 Huang et al.37 added a 1.8% slurry of PTFE directly to dosed samples and, the eYcacy of which is mirrored in the numerous eVorts targeting sample desolvation for pneumatic nebulization using by charring the sample at 450 °C, eVected decomposition of the PTFE to liberate fluorine which subsequently attacked the heated spray chambers with tandem condensers or membrane driers to remove moisture.This is particularly advantageous sample matrix and enhanced the volatility of the analyte. A variation of this approach was used to separate As39 and when applied to the determination of the rare earth elements, wherein multiple corrections for polyatomic oxides are other- Si40 selectively from sample matrices by the addition of an aliquot of 3% (m/v) NaF solution to a mixture of the sample wise necessitated.29,30 The same benefits can be more eYciently realized with the ETV.Additionally, reduced solvent load and H2SO4. The heat of the reaction was suYcient to volatilize AsF3 and SiF4 from the medium for transport to an ICP (for alters plasma chemistry in that electron density is lowered but the excitation temperatures are increased, less power is AES detection). Although the reaction was conducted in a separate PTFE reservoir, extrapolation of this approach to expended in vaporizing/dissociating water and the plasma ionization characteristics are improved, thereby enhancing ETV-ICP-MS is straightforward.Kumamaru and co-workers advantageously utilized the signal intensity.31 ETV as a thermochemical reactor to undertake in situ alkylation reactions.42–44 Following the dosing of the liquid sample Minimizing transport of matrix components. Apart from the elimination of solvent load, sample matrix load can also be and application of a drying stage to remove all water, the residue was reacted with a solution of ethylmagnesium bromide altered by one of two approaches: the sample matrix may be selectively vaporized in a separate step, prior to the vaporiz- in tetrahydrofuran to permit release of the volatile diethylberyllium at temperatures as low as 600 °C.42 Similar approaches ation and introduction of the analyte into the plasma or, in reverse fashion, the matrix is selectively removed from the also permitted the formation and volatilization of dibutylzinc at 250–450 °C43 following in situ butylation with butyllithium ETV following early (low temperature) transfer of the analyte to the plasma.In either case, it is often not suYcient to simply and the volatilization of ethylgallium at 600 °C44 by reaction with ethylmagnesium bromide. In each case, quantitative apply an appropriate thermal program, as the volatilities of the analyte and matrix are frequently not disparate enough to release of the analyte could be achieved, permitting its complete separation from the matrix.Low temperature vaporization accomplish significant separation eYciently. For this reason, chemical modifiers are often utilized which react with either (900 °C) of the 8-hydroxyquinolate complex of V45 and Cr (950 °C)46 was used to facilitate determination of these the analyte or the matrix components to create new compounds which enhance their volatility diVerences and permit more elements in rock, steel and aluminium samples following addition of 10 ml of a 0.2 M solution of reagent to the sample eYcient and selective vaporization of one of them.The ‘art’ of such applications is well developed and borrowed from the in the furnace. Although the above studies were conducted with an ETV-ICP-AES system, these approaches are clearly literature on ETAAS.32–34 Unfortunately, addition of modifiers is often plagued with the adventitious introduction of concomi- amenable to operation with ICP-MS.Indeed, Byrne et al.41 also reported on the in situ formation of a volatile 8-hydroxy- tant analyte impurities. These increase the analytical blank and degrade the detection limit and/or create new spectral quinolate complex of Cr for ETV-ICP-MS analysis which enhanced the limit of detection 20-fold compared with vaporiz- interferences via reaction of the relatively massive (mg scale) amounts of modifier with plasma gas species (i.e., generation ation without the reagent. This was ascribed to a lowering of the required vaporization temperature (to 1800 °C) which of argide species).35 J.Anal. At. Spectrom., 1999, 14, 785–791 787minimized the release of carbon into the ICP, thereby decreas- oxidative conversion is not completed, use of ascorbic acid alone results in the loss of a fraction of the total As at ing the intensity of the polyatomic 12C40Ar+ species interfering with the measurement of the principal Cr isotope. temperatures as low as 200 °C.Following completion of the oxidation step, ascorbic acid modifier was then added to promote early release of As and the sample vaporized at a Altering release of volatile elements. Separation of relatively volatile analytes from matrices of similar volatility presents a relatively low temperature of 500 °C (using a ramped heating of 300 °C s-1). Fig. 2 illustrates the recorded transient along diYcult challenge, as exemplified by the determination of elements such as Zn and Cd in seawater.The massive amount with that for 40Ar37Cl+. It is clear that nearly complete separation of the As from the matrix has now been achieved of salts that co-volatilize with these elements create not only excessive space charge problems, which virtually suppress the and there is insignificant ‘isobaric’ correction required for interference from 40Ar35Cl+. Total As in this sample was analyte ion signal,48,49 but can interfere with the measurement of other analytes (e.g., As) due to formation of polyatomic determined by the method of additions to yield 2.0±0.3 (n= 26) ng ml-1 as compared with 1.69±0.09 ng ml-1 obtained species (40Ar35Cl+). Two approaches to this problem arise: stabilization of the by hydride generation ETAAS following photo-oxidation of the sample.A limit of detection for As was estimated to be analyte with use of a classical modifier, such as reduced palladium, while attempting to volatilize the major fraction of 7.5 ng l-1 with a sample throughput of 15 per h.A further example of the utility of the ETV to function as matrix selectively, or formation of a more volatile complex of the analyte which permits its selective low temperature release. a thermochemical reactor can be illustrated by the direct determination of Cd in seawater.57 Fig. 3A shows transients An example of the former is the work of Gre�goire and Ballinas48 who utilized Pd and Mg to stabilize As while treating for release of 100 pg of Cd from a standard 1% (v/v) solution of HNO3 using a maximum power heating mode for the ETV the sample with NH4NO3 (to form NH4Cl ) in an eVort to volatilize as much chloride (in the form of NH4Cl ) as possible (approximate heating rate of 1800 °C s-1).Fig. 3B presents a trace for the release of 100 pg Cd in the presence of 0.2 mg at 1000 °C prior to the determination of As in seawater (to minimize 40Ar35Cl+ polyatomic interference). EDTA when using a ramped heating of the ETV (1000 °C s-1).The integrated intensities for traces A and B are, within Fig. 1 illustrates transient signals for As obtained for a similar situation encountered in our own laboratory. Despite experimental uncertainty, the same.ig. 3C presents a trace for the ramped atomization of 200 pg Cd without EDTA, use of an Ni modifier to enhance the thermal stability of As, and in addition to the introduction of NH4NO3 coupled with from which it is clear that the integrated intensity for the 111Cd is less than 1% of its response in the presence of EDTA.a pyrolysis temperature of 1100 °C, the signal for As is significantly suppressed in a neat sample of 20 ml of seawater Atomic absorption spectrometry was used to show that the Cd was likely released at approximately 700 °C in the form of (NASS-4 reference material ). Additionally, it is clear that complete removal of the chloride fraction of the matrix is a cold vapor from samples treated with EDTA.Chapple and Byrne49 adopted a diVerent approach to the unsuccessful, as the recorded 40Ar37Cl+ reveals that 75As would also be compromised by overlap with more abundant problem in that multiple additions of HNO3 were made to the sample in combination with pyrolysis at 1200 °C in an eVort 40Ar35Cl+. Based on earlier reports that use of EDTA or organic diacids, such as ascorbic or citric acid, is useful for to remove the salt matrix from seawater samples. Volatile elements were lost in the process and only Co, Cu, Mn, Ni facilitating early release of volatile elements from saline media,50–52 experiments were conducted to address the eYcacy and V could subsequently be quantitated.A further example of the utility of the ETV functioning as of this approach for the direct determination of As in seawater. Further, as it is well-known that environmental samples may a reactor for in situ sample processing is the thermal generation of a volatile complex of Se by reaction with added citric acid contain As in several diVerent chemical forms, i.e., monomethylated and dimethylated species, arsenocholine and arsenosu- to promote its quantitative and early release from a matrix of digested sediment.58 Atomic absorption spectrometry was sub- gars53–55 (arsenobetaine has not been reported in seawater, rather a halogenated betaine-like species), which may have sequently used to show that the Se volatilized as a molecular complex from the sample at temperatures below 600 °C (i.e., diVerent volatility and chemistry for interaction with Pd or Ni modifiers,56 it was deemed appropriate to ensure that all forms no atomic absorption was registered).Fig. 4 illustrates typical Se signal transients generated in the absence and presence of of As likely to be present in the sample be first converted to a common (inorganic) form. This was conveniently achieved added citric acid. Although sensitivity is enhanced more than 50-fold, the depression evident in the tail of the signal in the in situ by undertaking oxidation of the sample at 500 °C while admitting a 200 ml min-1 flow of air into the ETV.If this presence of citric acid is a consequence of the concurrent release of excess modifier from the ETV. This space charge interference, which likely occurs throughout the entire vaporiz- Fig. 1 Signal transients for As using ETV-ICP-MS sample introduc- Fig. 2 Signal profiles for As and background in NASS-4 seawater for tion.a 20 pg As standard, 300 °C pyrolysis temperature; b 20 ml sample of seawater (NASS-4, equivalent to approximately 40 pg As) spiked ETV sample introduction following air oxidation, addition of ascorbic acid, pyrolysis at 400 °C and ramped atomization (5 s) to 1400 °C, 2 s with 5 mg Ni and 10 mg NH4NO3 modifiers, 1100 °C pyrolysis temperature; and c trace for 40Ar37Cl+. read delay. a 75As; and b 40Ar37Cl+. 788 J. Anal. At. Spectrom., 1999, 14, 785–791desorbing from the graphite surface.A particularly powerful example of the attributes of this added dimensionality is the use of ETV for sample introduction into an ICP-TOF-MS instrument.24 The characteristic appearance temperatures of several elements (Cd, Sn and In) were used to separate the signals temporally when they would otherwise exhibit mutual (isobaric) interference. Attempting this by mass spectrometry alone would have demanded unattainable resolution with the system used (90–100 k).In situ speciation Answers to questions relating to toxicity, bioavailability and transport processes are highly dependent on an element’s form and can only be ascertained by acquiring quantitative species specific information.59 In addition to valence state speciation information, identification of organometallic species (of As, Se, Sn, Hg and Pb) has been most actively pursued. Although numerous tandem source approaches have been utilized relying on the chromatographic characteristics of the temporally- or volume-resolved response from atomic detectors to identify the species detected,59,60 little has been accomplished with use of the ETV.While admittedly the examples which follow could be, in some cases, more easily addressed with other techniques, they nonetheless provide evidence of the capabilities of the ETV in this area of study. Richner and Wunderli61 used the ETV to eVect a thermal separation of inorganic chlorine from its more volatile forms associated with polychlorinated biphenyls, enabling quantitation of the latter in waste oils by ICP-MS. Organochlorine could be removed from the sample aliquot by heating the ETV to 400 °C, whereas the inorganic fraction was removed by heating to 2650 °C. Discrimination of the inorganic fraction from total mercury present in biological tissue was reported by Willie et al.62 and Fig. 3 Signal profiles for 111Cd with ETV sample introduction, 350 °C provided an elegant example of the use of the ETV as a pyrolysis.A, 100 pg Cd with maximum power heating to 1900 °C; B, 100 pg Cd with 0.2 mg EDTA, ramped heating at 1000 °C s-1; and C, thermochemical reactor for in situ sample pretreatment and 200 pg Cd, no modifier, ramped heating at 1000 °C s-1. speciation. The integrity of the species was preserved by solubilization of the tissue with tetramethylammonium hydroxide (TMAH). For the determination of total mercury, sample aliquots were simply dried and vaporized into the plasma.For the selective determination of inorganic mercury, iodoacetic acid, sodium thiosulfate and acetic acid were added to the sample, cleaving the methylmercury from the tissue. Volatile methylmercury iodide so formed was released during the sample drying stage (120 °C), leaving only inorganic mercury to be quantitated by the method of additions. Other, less quantitative, speciation schemes for the determination of the various forms of mercury present in soils and sediments rely on simple thermal distillation or pyrolysis of a sample (2 mg) by heating at 0.5 °C s-1 and sweeping the evolved gas into a quartz tube for AAS detection.63,64 The same can be conveniently accomplished with the ETV, wherein the vapors are swept to the ICP torch.Calibration is achieved with use of standard reference soils. This is clearly likely to Fig. 4 Signal profiles for 78Se with (trace A, 50 pg Se) and without (trace B, 500 pg Se) citric acid.‘Pyrolysis’ temperature of 120 °C, provide a qualitative approach to this problem, as otherwise ramped (20 s) atomization to 2600 °C. the standard soils themselves must be certified for speciation content. Arpadjan and Krivan65 quantitatively achieved in situ ation event, noticeably compromised the 77Se582Se ratio but separation of CrIII from CrVI using the heated graphite furnace permitted the determination of ultra-trace concentration to initiate reaction of samples (water and urine) with a mixture (50 ng g-1) of Se in a sediment by the method of additions.58 of TMAH, methanol, sodium acetate and trifluoroacetylacetone Apart from the direct minimization of some polyatomic (TFA).The TFA complex of CrIII was removed from the cell interferences by judicious use of thermal programming and in by heating to 400–1200 °C (300 s heating at 400 oC), after situ thermochemical reactions with suitable reagents, as which the CrVI was volatilized by heating to 2600 °C.described above, the ETV also oVers the advantage of an Extrapolation of the procedure to ETV-ICP-MS detection added dimensionality to the analysis. Temporal release of remains to be verified in light of the problems which arise analytes from the surface of the ETV can be used to advantage to eVect a chromatographic separation of the various species from 52ArC+. However, the methodology reported by Byrne J. Anal. At.Spectrom., 1999, 14, 785–791 789et al.41 and Tao and Kumamaru46 may likely be utilized to Although the principal application has been directed to the in situ trapping of hydride forming elements,74 a number of circumvent this problem by release of CrVI at low temperature. A similar approach to the ultra-trace diVerential determi- volatile trace element compounds can be targeted for such application, including phosphorus,75 sulfur76 and several trans- nation of AsV and AsIII in aqueous environmental samples was reported by Chen et al.66 When the graphite surface of the ition metals, i.e., Cd,77 Cu78, Ni79 and Pb80.Additionally, the ETV may be conveniently used as a preconcentration cell for ETV was modified by high temperature impregnation with zirconium, it was possible to volatilize AsIII selectively in the the collection of aerosol particulates81 using electrostatic deposition techniques. This approach has also been successfully form of the chloride by addition of 9 M HCl to the sample and heating to 400 °C.The AsV was retained in the tube to demonstrated for the collection of the volatile hydrides of As, Se and Sb.13 temperatures up to 1400 °C by interaction with the surface. The method of additions was used for quantitation. At this The ETV may also function as the cathode of an electrochemical cell, thereby permitting the deposition of electroactive temperature, most of the chloride has been removed from the furnace and spectral interference from 40Ar35Cl+ should be analytes from solution onto its surface.82 In situ matrix elimination for aqueous samples, with potentially quantitative elec- negligible.trodeposition of trace elements within 60 s using 20–40 mA In situ sample preparation currents, can be achieved. As an example, the deposition process was automated to include application of a modifier The ETV has also been used as an eYcient medium for solid reagent using a Pt/Ir autosampler (anode) and, in conjunction sample digestion.The relatively inert, high purity graphite with a rinse cycle for removal of the electrolysed sample, substrate and programmable temperature may be used to >99.5% of a 0.5 M NaCl matrix was eliminated. The advanprocess samples in situ more eYciently than oV-line bulk tage of this approach to minimization of space charge matrix processing. An example of this is reported by Okamoto et al.,67 interferences is clear, as is extrapolation of the technique to wherein a weighed, powdered botanical sample was mixed the potential speciation of the diVerent electroactive forms with ammonium phosphate (modifier) and placed into an ETV of elements.device. TMAH was added to eVect an in situ wet digestion at 150 °C. Ashing of the sample at 1000 °C removed the bulk of Conclusion the matrix and the Cd analyte was then volatilized into a plasma by heating to 2500 °C. The most remarkable feature The full potential of ETV sample introduction for tandem of the technique was that both sample decomposition and source ICP-MS remains to be explored and exploited.The vaporization were accomplished using the same vessel. microenvironment of the ETV is convenient for use of the cell A further example of the use of the ETV for in situ sample as a thermochemical reactor for the pretreatment of samples decomposition is the simple oxygen ashing of biological mate- which cannot be easily handled by other means in the laborarials that can be achieved during slurry or solid sampling of tory.The ease with which solid (as slurry), liquid and gaseous tissues, wherein the organic matrix can be eliminated by reagents can be admixed with the sample in this graphite thermally treating the sample at 600–800 °C in the presence of thermochemical reactor is attractive. The precise programming air.68,69 Any oxidative destruction of the pyrolytic graphite of the temperature in this environment opens the door for coating that may arise as a result of using the higher tempera- sample pretreatment with the aim of synthesis of volatile tures likely does not impact on the ICP-MS response as analyte complexes, elimination of matrix components, selective significantly as with ETAAS since only volatilization of the vaporization of analyte species and simple decomposition of sample is sought.More recently,70 it has been suggested that complex materials.Each of these scenarios serves to enhance this process may also facilitate analyte transport to the ICP the performance of the analytical technique by minimizing as a consequence of reproducible formation of carbonaceous sample preparation time, reducing or removing matrix-induced aerosol species (from partial oxidative destruction of the tube polyatomic interferences and space charge eVects and enhancsurface) serving as transport nuclei. ing the relative limit of detection.Clearly, optimum benefit from these accomplishments will accrue from use of TOF In situ preconcentration instrumentation to take advantage of the ‘simultaneous’ registration of multielement information. It should be evident from The ETV functions eYciently for handling the eZuents (conthe foregoing that ETV satisfies more than just a niche end centrates) from flow injection manifolds71,72 and is thus able use for relatively simple and eYcient transfer of analyte to the to embrace all of the same advantages which accrue when plasma.ETV should not be viewed as competitive to other these systems are interfaced to any atomic spectrometric high eYciency microsample liquid introduction techniques, or detection system. In this respect, the ETV serves in a passive in relation to laser ablation for solid sampling devices or mode for tandem source sample introduction. hyphenated chromatographic approaches for speciation, but Apart from this obvious application, the ETV itself can rather as a complementary methodology in the arsenal of function in a more proactive mode, wherein it may serve as problem solving approaches.an eYcient preconcentration cell for liquids and gases to enhance relative detection power. The simplest approach in this direction is the use of spray deposition for sample dosing References into the ETV.73 Typically limited to 20–50 ml sample aliquots 1 R. Sturgeon, Fresenius’ J. Anal. Chem., 1996, 355, 425.when pipetted, this volume may be eVectively increased to 2 V. Karanassios and G. Horlick, Spectrochim. Acta Rev., 1990, several hundred ml by spray deposition of the sample directly 13, 89. into a preheated ETV (typically 160 °C) using a standard 3 Spectrochim. Acta, Part B, Analytical Atomic Spectrometry: From pneumatic nebulizer wherein the aerosol is simultaneously Furnace to Laser, 1989, 44, 1209–1403. 4 J. M. Carey and J. A. Caruso, CRC Crit. Rev. Anal. Chem., 1992, desolvated and adsorbed onto the wall of the ETV.Relative 23, 397. detection limits can be enhanced 10-fold with ideal samples. 5 D.C.Gre� goire, Can. J. Spectrosc., 1997, 42, 1. Similar in concept is the use of the ETV to sequester the 6 W. Frech and S. Jonsson, Spectrochim. Acta, Part B, 1982, 37, volatile forms of elements generated in chemical manifolds, 1021. thereby providing enhancements in relative detection limits by 7 T. Kantor, Spectrochim. Acta, Part B, 1988, 43, 1299.factors in excess of 1000-fold compared with solution phase 8 M. W. Borer and G. M. Hieftje, Spectrochim. Acta Rev., 1991, 14, 463. analysis. This methodology has recently been reviewed.11 790 J. Anal. At. Spectrom., 1999, 14, 785–7919 A. Montaser, M. G. Minnich, J. A. McLean and H. Liu, in 47 T. Kantor, Fresenius’ J. Anal. Chem., 1996, 355, 606. 48 D. C. Gre�goire and M. L. Ballinas, Spectrochim. Acta, Part B, Inductively Coupled Plasma Mass Spectrometry, ed. A. Montaser, Wiley-VCH, New York, 1998, pp. 83–264. 1997, 52, 75. 49 G. Chapple and J. P. Byrne, J. Anal. At. Srom., 1996, 11, 549. 10 M. Hornung and V. Krivan, Anal. Chem., 1998, 70, 3444. 11 H. Matusiewicz and R. E. Sturgeon, Spectrochim. Acta, Part B, 50 R. Guevremont, R. E. Sturgeon and S. S. Berman, Anal. Chim. Acta, 1980, 115, 163. 1996, 51, 377. 12 T. Buchkamp, T. G. Hermann, presented at the 3rd European 51 R. Guevremont, Anal. Chem., 1980, 52, 1574. 52 R. Guevremont, Anal. Chem., 1981, 53, 911.Furnace Symposium, Prague, Czech Republic, June 14–18, 1998, paper O6. 53 R. E. Sturgeon, K. W. M. Siu, S. N. Willie and S. S. Berman, Analyst, 1989, 114, 1393. 13 R. M. Camero and R. E. Sturgeon, Spectrochim. Acta, Part B, 1999, 54. 54 M. Ochsenkuhn-Petropulu, K.-M. Ochsenkuhn, I. Milonas and G. Parissakis, Can. J. Spectrosc., 1994, 40, 61. 14 L. Moens, P. Verrept, S. Boonen, F. Vanhaecke and R. Dams, Spectrochim. Acta, Part B, 1995, 50, 463. 55 J. J. Corr and E.H. Larsen, J. Anal. At. Spectrom., 1996, 11, 1215. 56 V. I. Slaveykova, F. Rastegar and M. J. F. Leroy, J. Anal. At. 15 N. J. Miller-Ihli, Spectrochim. Acta, Part B, 1995, 50, 477. 16 S. Boonen, F. Vanhaecke, L. Moens and R. Dams, Spectrochim. Spectrom., 1996, 11, 997. 57 W. H. Lam, R. E. Sturgeon and J. W. McLaren, presented at the Acta, Part B, 1996, 51, 271. 17 H. Baumann, Fresenius’ J. Anal. Chem., 1992, 342, 907. 1998 Winter Conference on Plasma Spectrochemistry, Scottsdale, Arizona, January 5–10, 1998, paper 61. 18 S. A. Darke and J. F. Tyson, Microchem. J., 1994, 50, 310. 19 U. Voellkopf, M. Paul and E. R. Denoyer, Fresenius’ J. Anal. 58 W. H. Lam, R. E. Sturgeon and J. W. McLaren, Spectrochim. Acta, Part B, 1999, 54. Chem., 1992, 342, 917. 20 S.-F. Chen and S.-J. Jiang, J. Anal. At. Spectrom., 1998, 13, 673. 59 R. Lobinski and Z. Marczenko, Spectrochemical Trace Analysis for Metals and Metalloids, Elsevier, Amsterdam, 1997. 21 C. M. Sparks, J.A. Holcombe and T. L. Pinkston, Appl. Spectrosc., 1996, 50, 86. 60 R. E. Sturgeon, J. Anal. At. Spectrom., 1998, 13, 351. 61 P. Richner and S. Wunderli, J. Anal. At. Spectrom., 1993, 8, 45. 22 U. Schafer and V. Krivan, Anal. Chem., 1998, 70, 482. 23 D. C. Gre�goire and R. E. Sturgeon, Spectrochim. Acta, Part B, 62 S. N. Willie, D. C. Gre�goire and R. E. Sturgeon, Analyst, 1997, 122, 751. 1999, 54. 24 P. P. Mahoney, S. J. Ray and G. M. Hieftje, Appl. Spectrosc., 63 G. Bombach, K.Bombach and W. Klemm, Fresenius’ J. Anal. Chem., 1994, 350, 18. 1997, 51, 16A. 25 D. E. Nixon, V. A. Fassel and R. N. Kniseley, Anal. Chem., 1974, 64 H. Biester and G. Nehrke, Fresenius’ J. Anal. Chem., 1997, 358, 446. 46, 210. 26 A. L. Gray and A. R. Date, Analyst, 1983, 108, 1033. 65 S. Arpadjan and V. Krivan, Anal. Chem., 1986, 56, 2611. 66 Y. Chen, W. Qi, J. Cao and M. Chang, J. Anal. At. Spectrom., 27 H. Matusiewicz, J. Anal. At. Spectrom., 1986, 1, 171. 28 A. Kh. Gilmutdinov, B.Radziuk, M. Sperling and B. Welz, 1993, 8, 379. 67 Y. Okamoto, N. Takagaki, T. Fujiwara and T. Kumamaru, Anal. Spectrochim. Acta, Part B, 1996, 51, 1023. Commun., 1997, 34, 283. 29 M. P. Field and R. M. Sherrell, Anal. Chem., 1998, 70, 4480. 68 D. K. Eaton and J. A. Holcombe, Anal. Chem., 1983, 55, 946. 30 G. Horlick and A. Montaser, in Inductively Coupled Plasma Mass 69 R. W. Fonseca, N. J. Miller-Ihli, C. Sparks, J. A. Holcombe and Spectrometry, ed. A. Montaser, Wiley-VCH, New York, 1998, B. Shaver, Appl. Spectrosc. 1997, 51, 1800. pp. 503–588. 70 V. Majidi and N. Miller-Ihli, Spectrochim. Acta, Part B, 1998, 31 V. Karanassios and G. Horlick, Spectrochim. Acta, Part B, 1989, 53, 965. 44, 1387. 71 J. L. Burguera and M. Burguera, Analyst, 1998, 123, 561. 32 D. L. Styris and D. A. Redfield, Spectrochim. Acta Rev., 1993, 72 L. C. Azeredo, R. E. Sturgeon and A. J. Curtius, Spectrochim. 15, 71. Acta, Part B, 1993, 48, 91. 33 D. L. Tsalev and V. I. Slaveykova, J. Anal. At. Spectrom., 1992, 73 J.-F. Alary and E. D. Salin, Spectrochim. Acta, Part B, 1995, 7, 147. 50, 405. 34 B. Welz, Atomic Absorption Spectrometry, VCH, Weinheim, 1985. 74 R. E. Sturgeon and D. C. Gre�goire, Spectrochim. Acta, Part B, 35 D. C. Gre�goire and R. E. Sturgeon, Spectrochim. Acta, Part B, 1994, 49, 1335. 1993, 48, 1347. 75 K. Fujiwara, M. A. Mignardi, G. Petrucci, B. W. Smith and J. D. 36 J. P. Matousek and H. K. Powell, Spectrochim. Acta, Part B, 1986, Winefordner, Spectrosc. Lett., 1989, 22, 1125. 41, 1347. 76 A. Syty, Anal. Chem., 1979, 51, 911. 37 M. Huang, Z. Jiang and Y. Zeng, J. Anal. At. Spectrom., 1991, 77 H. Matusiewicz, M. Kopras and R. E. Sturgeon, Analyst, 1997, 6, 221. 122, 331. 38 K. C. Ng and J. A. Caruso, Analyst, 1983, 108, 476. 78 R. E. Sturgeon, J. Liu, V. J. Boyko and V. Luong, Anal. Chem., 39 A. Lopez-Molinero, M. Benito, Y. Aznar, A. Villareal and J. R. 1996, 68, 1883. Castillo, J. Anal. At. Spectrom., 1998, 13, 215. 79 R. E. Sturgeon, S. N. Willie and S. S. Berman, J. Anal. At. 40 A. L. Molinero, A. Morales, A. Villareal and J. R. Castillo, Spectrom., 1989, 4, 443. Fresenius’ J. Anal. Chem., 1997, 358, 603. 80 R. E. Sturgeon, S. N. Willie and S. S. Berman, Anal. Chem., 1989, 41 J. P. Byrne, R. McIntyre, M. E. Benyounes, D. C. Gre�goire and 61, 1867. C. L. Chakrabarti, Can. J. Spectrosc., 1997, 42, 95. 81 G. Torsi, P. Reschiglian, M. Lippolis and A. Toschi, Microchem. 42 S. Tao and T. Kumamaru, Anal. Chim. Acta, 1994, 292, 1. J., 1996, 53, 437. 43 S. Tao, Y. Okamoto and T. Kumamaru, Anal. Sci., 1995, 11, 319. 82 J. P. Matousek and H. K. J. Powell, Spectrochim. Acta, Part B, 44 T. Kumamaru, S. Tao, M. Uchida and Y. Okamoto, Anal. Lett., 1995, 50, 857. 1994, 27, 2331. 45 S. Tao and T. Kumamaru, Appl. Spectrosc., 1996, 50, 785. 46 S. Tao and T. Kumamaru, J. Anal. At. Spectrom., 1996, 11, 111. Paper 8/09460H J. Anal. At. Spectrom., 1999, 14, 785&nda

ISSN:0267-9477    

DOI:10.1039/a809460h

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Correction for volatility diVerences between organic sample analytes and standards in organic solutions analyzed by inductively coupled plasma-atomic emission and mass spectrometry† Assad S. Al-Ammar, Rajesh K. Gupta and Ramon M. Barnes* Department of Chemistry, Lederle Graduate Research Center Towers, University of Massachusetts, Box 34510, Amherst, MA 01003–4510, USA Received 29th October 1998, Accepted 26th February 1999 A novel technique was developed to correct for the error in inductively coupled plasma atomic emission and mass spectrometric measurements arising from a diVerence in volatility between the sample analyte compounds and standards.The technique is based on the measurement of the analyte signal at two spray-chamber temperatures. A volatility correction factor is then estimated from a linear correlation between the reciprocal of a correction factor and the relative change in intensity resulting from measurements at two spray-chamber temperatures.Tests with organosilicon and organochlorine compounds demonstrate a significant decrease (from 2 to 30 times) in error after correction. The technique requires no prior knowledge of the chemical structure of the analyte. The determination of trace elemental impurities by inductively Theoretical coupled plasma atomic emission spectroscopy (ICP-AES) and Development inductively coupled plasma mass spectrometry (ICP-MS) in some petroleum products and metal alkyls, that are used for The theory developed is based on two equations: the Clausius– semiconductor materials manufacturing, is subject to a high Clapeyron and Browner equations.In the Clausius–Clapeyron determination error. This error arises from the diVerences in equation, the partial vapor pressure, P, of the substance is volatility between the sample analyte compound and the related to the absolute temperature: standard used for calibration. When several organometallic log P=-(0.05223a/T )+b (1) compounds of an element with significantly diVerent boiling points are measured by ICP-AES or ICP-MS, the vapor where a and b are constants related to the compound introduced into the discharge from the spray chamber can volatility.7 become selectively enriched in the more volatile compound.1 The second equation, derived by Browner et al.,8–10 relates The signal for the volatile species will be enhanced relative to the rates of evaporation of a pure substance in a form of the non-volatile species.Hence compounds with lower boiling- liquid droplets at temperature, T, to its physical parameters. points produce a higher signal per mole of analyte than do Thus, rate of evaporation=[8prDvM2s/r(RT )2]P where r is species with higher boiling-points. Accordingly, when measur- the droplet radius, Dv is the vapor diVusion coeYcient, M is ing a sample solution containing analyte compounds with the molecular mass, s is the surface tension, P is the partial volatilities diVerent from that of the standard used, the analysis pressure of the material at temperature T , r is the density and result will suVer from an error as large as several thousand R is the gas constant. This equation was originally derived to per cent.1 On the other hand, some analytes with suYciently describe the liquid droplet evaporation rate in an ICP-AES high volatility yield very high sensitivity enhancement that spray chamber. The total evaporation rate is the sum of this may be as high as 30 times.1 Several approaches have been equation over all the drops inside the spray chamber. For applied to eliminate volatility enhancement error.These present development the most important feature of this equainclude exponential dilution,1 sample decomposition1,2 and tion is the direct proportionality between the compound direct injection nebulizer (DIN) sample introduction.3–6 These evaporation rate and its partial vapor pressure.techniques, however, also eliminate any advantage gained from The terms r, Dv, M, s and r can be grouped together in one the sensitivity enhancement. A disadvantage of the DIN is term, which could be regarded as constant, since the solvent plasma cooling that occurs while an organic solvent is nebul- used to prepare the sample and standard is the same. Thus, the ized. This degrades the detection limits for most elements by Browner equation can be transformed to the following: as much as an order of magnitude compared with conventional rate of evaporation (moles per unit time)=k(P/T 2) (2) sample introduction methods.6 The aim of this investigation was to develop a simple, The value of the constant k depends on the nature of the solvent, while P represents the vapor pressure of the solute eYcient technique to correct for volatility enhancement.The technique should require no prior knowledge of the chemical (i.e., the analyte compound). kP/T 2 is the number of moles of analyte as vapor reaching the plasma per unit time.Since structure of the volatile analyte. The approach also should maintain the improved detection limit resulting from the high the intensity of the ICP-AES analyte signal, I, increases linearly with the total number of moles of the analyte reaching the volatility of low boiling-point analyte compounds. plasma per unit time, eqn. (2) can be extended to reflect I: I=SM+Sk(P/T 2) (3) †Presented in part at the 5th Rio Symposium on Atomic Spectroscopy, where S is a constant that relates the analyte intensity to the Cancu� n, Mexico, October 4–10, 1998, and the 1999 European Winter number of moles of analyte reaching the plasma per unit time. Conference on Plasma Spectrochemistry, Pau, France, January 10–15, 1999.M is the number of moles of the analyte reaching the plasma J. Anal. At. Spectrom., 1999, 14, 793–799 793per unit time as part of the solvent droplets. If the same solution are at exactly the same temperature.In an actual experiment, only the spray chamber walls are chilled to 0 °C. analyte solution is measured at two diVerent temperatures (i.e., temperature T1 for which the spray chamber is cooled to The sample temperature, however, is equal to room temperature while being nebulized. In addition, a self-cooling eVect of and maintained at 0 °C and temperature T2 for which the spray chamber is maintained at room temperature), then the the nebulized sample droplets results from solvent vaporization. This eVect cools the small droplets to a larger extent change in intensity, DI, can be expressed as than large droplets.Accordingly, the nebulized sample aerosol DI=I2-I1=Sk(P2/T22-P1/T12) (4) may have a temperature gradient. Furthermore, the eVect of where the subscripts 1 and 2 indicate measurements at T1 and changing the nebulized sample’s viscosity resulting from the T2, respectively. In deriving eqn.(4), it is assumed that M is change of spray chamber temperature was not considered. The not changed appreciably when the spray chamber is cooled solvent viscosity has pronounced eVect on the volatility of from T2 to T1, because the solvents used in ICP-AES usually the solute. have low volatility. The correlation between the correction factor, I10/I1, and Also, since DI/I1 should be determined experimentally to construct an accurate scheme for volatility correction based on eqn.(14). DP=P2-P1, (5) In this study two compound classes were tested. Organosilicon then from eqns. (4) and (5): compounds with a wide volatility range were dissolved (10 mg Si L-1) in three solvents (m-xylene, dec-1-ene and hexan-1-ol ). DI=Sk(DP/T22+P1/T22-P1/T12) (6) Aqueous solutions also were prepared from organochlorine Taking the derivative of P with respect to T in eqn. (1): compounds with diVerent volatility (5 mg Cl L-1). The composition and the details of preparation are discussed in the dP/dT=0.12aP/T 2 (7) Experimental section.The silicon compound containing soluwhen the change in temperature is not too large (e.g., 0 to tions were measured by ICP-AES and the chloro compound 25 °C), then solutions by ICP-MS. The details of measurements are reported in the experimental section. These silicon and chlorine DP=(dP1/dT )(T2-T1) (8) solutions were measured at two diVerent spray chamber tem-rom eqns. (7) and (8): peratures, 0 and 25 °C.From these measurements, the inverse of the correction factor, 1/(I10/I1), was plotted against DI/I1 DP=(0.12aP1/T12)(T2-T1) (9) as indicated by eqn. (14). For all the cases studied, a linear From eqns. (6) and (9): regression was fitted. Negative slopes for dec-1-ene and hexan- 1-ol and positive slopes with water and m-xylene solvents were DI=SkP1[0.12a(T2-T1)/T12T22+1/T22-1/T12) (10) obtained (Fig. 1–4). If the correction factor instead of its From eqns. (3) and (10): inverse were plotted against DI/I1, then a curved line was found.DI=(I1-SM)[0.12a(T2-T1)/T22+T12/T22-1] (11) or Applications I1-SM=DI/[0.12a(T2-T1)/T22+T12/T22-1] Considering the experimentally obtained correlation between the inverse of the correction factor, 1/(I10/I1), and DI/I1 Let I10 be the intensity expected at temperature T1 when (Fig. 1–4), an analytical scheme was developed to quantify the analyte compound is completely non-volatile (i.e., the elements that exist as compounds having widely diVerent analyte is transported to the plasma only as liquid droplets).volatility with reference to a single standard containing these I10 represents the value of I1 after correction by eliminating elements. In this approach the standard contains non-volatile the transport of the analyte to the plasma as vapor [i.e., I10= analyte compounds. A non-volatile analyte standard is used, SM according to eqn. (3)]. because all commercially available trace element organometal- Dividing both sides of eqn.(11) by I10 and rearranging lic calibration standards in organic matrices are prepared with yields I1/I10=1+DI/{I10[0.12a(T2-T1)/T22+T12/T22-1]} (12) With simple manipulations I1 can be shown to be directly proportional to I10, thus, I10=hI1 (13) where h is the proportionality constant. From eqns. (12) and (13), the final relationship is derived: I1/I10=(DI/I1)/{h[0.12a(T2-T1)/T22+T12/T22-1]}+1 (14) Volatility correction A scheme to correct for volatility during nebulization is based on the theoretical relationship given by eqn.(14). This equation indicates the possibility of calculating a correction factor, Fig. 1 Volatility correction using control solutions 1, 2, 3 and 4 and I10/I1, from the experimentally measured intensity ratio DI/I1. standard solution in xylene. The inverse of correction factor I1/I10 is However, construction of an accurate volatility scheme based plotted as the ordinate and the % increase in signal, (DI/I1)×100, is on eqn.(14) is limited, because this equation does not represent plotted as the abscissa. Slope=0.71±0.12, intercept=-0.6748, r2= precisely the actual experimental conditions. In deriving eqn. 0.9444, standard error of slope=0.09948, 95% confidence interval of mean±4.58. (14) we assumed that both the spray chamber and the analyzed 794 J. Anal. At. Spectrom., 1999, 14, 793–799depends on the magnitude of the attractive force between the solvent and the solute molecule.The analysis protocol is based on the use of one calibration and several control solutions in a solvent common to the sample. These solutions contain exactly the same analyte concentration. However, analyte compounds with significantly diVerent volatility must be used to prepare the control solutions, while the calibration solution is prepared using a non-volatile analyte compound. The control, sample and control solutions are measured at two diVerent spray chamber temperatures.These measurements are used to calculate DI/I1 (or (DI/I1)×100) and I1/I10 for each control solution. I10 is taken to be the measurement of the calibration solution at temperature T1. The relative DI/I1 instead of the absolute DI value is preferred, because it depends only on the temperature change and can be applied to samples of unknown concentration. DI depends on both concentration and temperature change. The calculated Fig. 2 Volatility correction using control solutions 1, 2, 3, 4 and 5 in control solution I1/I10 values are plotted against the DI/I1 dec-1-ene.Slope=-9.4±1.6, intercept=7.6792, r2=0.9113, standard values to construct a linear regression function. This plot error of slope=0.01685, 95% confidence interval of mean±1.27. represents a linear relationship between the inverse of the correction factor 1/(I10/I1) and the relative change in intensity (DI/I1)×100 accompanying the change in the spray chamber temperature from T1 to T2.This relative change in intensity furnishes a quantitative measure of the analyte compound volatility even if its structure is unknown. The relative change in sample intensity is measured, and this value is then used to obtain the correction factor I10/I1 from the constructed correction curve. The correction factor is then multiplied by the measured sample intensity to transform it to I10. This value corresponds to the predicted intensity corrected for the volatility diVerence of the analyte compound in the sample.Thus, this correction transforms I1 to a value expected for a non-volatile compound. Finally, the corrected intensity I10 is used to calculate the correct concentration by reference to the calibration solution. Compound mixtures The above derivation for a single compound sample analyte Fig. 3 Volatility correction using control solutions 1, 2, 3, 4 and 5 in hexan-1-ol. Slope=-16.5±2.8, intercept=7.5864, r2=0.9186, stan- can be extended to samples composed of a mixture of several dard error of slope=0.02847, 95% confidence interval of mean±1.12.compounds with diVerent volatilities. Moreover, the proposed scheme applies without modification. Let I11, I12, …, I1 n be the intensities (measured at T1) for the first, second, …, nth compounds. In practice, one could measure only the total intensity, I1total, which is equal to the sum of I11, I12, …, I1 n. Let DI1, DI2, …, DIn be the diVerences in intensities when the same sample is measured at two temperatures T1 and T2. One could also measure only the total diVerence, DItotal, which is equal to the sum of DI1, DI2, …, DIn.From eqn. (11) we obtain I11-SM1+I12-SM2+…+I1 n-SMn =[1/(0.12a(T2-T1)/T22+T12/T22-1)] ×(DI1+DI2+…+DIn) (15) In eqn. (15), a normalized value of the constant a was used that has a value in the range ai<a<ai¾, where ai and ai¾ represent the constant a for the compound with highest and the lowest volatility in the mixture, respectively.Eqn. (15) can be condensed by dividing both sides by I10: Fig. 4 Volatility correction using control solutions 1, 2, 3 and 4 and I1total/I10=1+DItotal/{I10[0.12a(T2-T1)/T22+T12/T22-1]} standard in water. Slope=31.9±5.4, intercept=-0.1157, r2=0.9135, (16) standard error of slope=0.04929, 95% confidence interval of mean±1.36. where (SM1+SM2+…+SMn)/I10=1. From eqns. (13)and (16): non-volatile organic analyte compounds such as salts of car- I1total/I10=DItotal/{I1h[0.12a(T2-T1)/T22+T12/T22-1]}+1 boxylic and sulfonic acids.Furthermore, in this scheme the (17) samples and calibration solution must be prepared using the same solvent. This requirement exists because the constant a Eqn. (17) has exactly the same form as eqn. (14). Therefore, the proposed scheme also could be applied without in eqn. (14), which is a measure of the analyte volatility, also J. Anal. At. Spectrom., 1999, 14, 793–799 795modification to a mixture of several compounds of the same based on the optimization reported in manufacturer’s instrument manuals. analyte.Furthermore, the correction curve derived from the control solution of one element can be applied to correct for Reagents and calibration solutions the volatility of other elements. This can be inferred from the observation that eqns. (14) and (17) contain variables and Two compound classes were tested. Organosilicon compounds constants that depend only on the volatility property of the with a wide volatility range were dissolved (10 mg Si L-1) in compound.They are independent of the chemical composition, three solvents (m-xylene, dec-1-ene and hexan-1-ol ). Aqueous the analyte identity and the spectral line or isotope used for solutions were prepared from organochlorine compounds with measurement by ICP-AES or ICP-MS, respectively. The diVerent volatility (5 mg Cl L-1). dependence of the sensitivity of the ICP-AES and ICP-MS measurement on the nature of the analyte and the spectral Experiment 1 (solvent, m-xylene; ICP-AES).Silicon comline or the isotope used is eliminated by the use of relative pounds (99% purity, Aldrich, Milwaukee, WI, USA) were units for the correction factor, I0/I1, and the change in dissolved in m-xylene (anhydrous grade, bp 139 °C, 99% purity, intensity, DI/I1. This property of using a single correction Aldrich) to prepare three types of test solutions: samples, curve that is derived from control solutions of one element calibration solutions and control solutions.The anhydrous mmakes the proposed scheme very practical. xylene was pre-treated with octadecyltrichlorosilane (Aldrich) to remove any residual moisture. Vapor loading Four control solutions, each containing 10 mg Si L-1, were prepared from four organosilicon compounds. Control solu- An error may arise when applying the proposed scheme using tion 1 was prepared from bis(trimethylsilyl )methane (bp a solvent with suYcient volatility to produce an appreciable 132 °C), control solution 2 from dichlorodimethylsilane (bp increase in vapor loading at temperature T2 compared with 70.5 °C), control solution 3 from chlorotrimethylsilane temperature T1.The higher vapor loading at temperature T2 (bp 58 °C) and control solution 4 from tetramethylsilane decreases the magnitude of signal enhancement of the analyte, (bp 26 °C). One calibration solution containing 10 mg Si L-1 DI.This eVect can be corrected by multiplying the analyte was prepared from aminopropyltriethoxysilane (bp 217 °C). signal at T2 by the ratio of signals from the standard at the Four synthetic sample solutions listed in Table 2 also were two temperatures T1 and T2 (SignalT1/SignalT2). This ratio prepared. Pure m-xylene was measured as the blank. reflects only the eVect of changing the solvent vapor loading, since the analyte compound in the standard is chosen to be Experiments 2 and 3 (solvents, dec-1-ene and hexan-1-ol; ICP- non-volatile.AES). Five control solutions each containing 10 mg Si L-1 were prepared from five diVerent organosilicon compounds in Experimental two diVerent solvents, dec-1-ene (98% purity, bp 181 °C, Aldrich) and hexan-1-ol (97% purity, bp 153 °C, Aldrich). Instrumentation Control solution 1 was prepared from bis(trimethylsilyl )methane, solution 2 from tetraethylsilane (bp 153 °C), solution 3 A commercial ICP-AES system (Optima 3000, Perkin-Elmer, Norwalk, CT, USA) was used for all experiments with organ- from hexamethyldisilane (bp 112 °C), solution 4 from trimethoxymethylsilane (bp 103 °C) and solution 5 from tetra- osilicon compounds prepared in m-xylene, dec-1-ene and hexan-1-ol solvents.The silicon emission was measured at Si methylsilane. A calibration solution containing 10 g Si L-1 was prepared from aminopropyltriethoxysilane in hexan-1-ol. I 212.412 nm with background correction. A commercial ICP-MS system (Spectromass 2000, Spectro Analytical A calibration solution containing 10 mg Si L-1 was prepared from octadecyltrichlorosilane (bp 223 °C) in dec-1-ene.Instruments, Fitchburg, MA, USA) was used for all experiments with organochlorine compounds prepared in water. The Five synthetic sample solutions listed in Tables 2 and 3 were prepared in dec-1-ene as the solvent. Samples 1–3 comprised ICP-MS 35Cl and 34S signals were recorded at m/z 34 and 35 for the chlorine and sulfur measurements, respectively.The single pure compounds and samples 4 and 5 contained a mixture of 21.3 mg Si L-1 from trimethoxymethylsilane and experimental operating parameters (Table 1) were selected Table 1 Operating conditions for ICP-AES and ICP-MS analyses ICP-AES ICP-MS Optima 3000 prototype Spectromass 2000 ICP system Rf power/kW 1.3 1.35 Frequency (free running)/MHz 40 27 ICP torch Type 2 quartz slotted extension Fassel type Torch injector Ceramic alumina Quartz Outer argon flow rate/Lmin-1 15 15 Intermediate argon flow rate/L min-1 2.0 1.5 Central argon flow rate/L min-1 0.8 0.95 Nebulizer Concentric glass (Glass Expansion, Concentric glass (Glass Expansion, Hawthorne, Victoria, Australia, Model Model 38493) 38493) Sample pump rate/mL min-1 0.8 0.8 Pump tubing Viton, orange–orange (id 0.035 in) Tygon (id 0.040 in) Spray chamber Glass Scott double-pass coolant jacketed Glass Scott double-pass coolant jacketed (Spectro Analytical Instruments) (Spectro Analytical Instruments) Spray chamber temperatures/°C 0,22 0,22 Integration time (auto)/s 10 5 Background correction/nm ±0.04 — Drain Pumped Pumped Detector voltage/V — 2250 796 J.Anal. At. Spectrom., 1999, 14, 793–7992.5 mg Si L-1 from tetramethylsilane, and of 15.9 mg Si L-1 In this test organosilicon compounds with wide range of as hexamethyldisilane and 2.3 mg Si L-1 as tetramethylsilane, volatilities (bp 26–223 °C) and dissimilar chemical structures respectively. Five sample solutions listed in Tables 2 and 3 (i.e., halogen, ethyl, methyl, methoxy, etc., functional groups) also were prepared in hexan-1-ol.Sample 5 contained a mixture were used. Also, organochlorine compounds with a wide range of 15.0 mg Si L-1 as trimethoxymethylsilane and 1.64 mg Si of volatilities (from CH2Cl2 with bp 40 °C to the non-volatile L-1 as tetramethylsilane. Pure dec-1-ene and pure hexan-1-ol NH4Cl ) were examined.The solvents used, m-xylene, hexanwere measured as blanks. 1-ol, dec-1-ene, and water, are very distinct in chemical structure and physical properties, such as viscosity, surface tension, Experiment 4 (solvents, propan-2-ol–water; ICP-MS). Three boiling-point, and polarity. Accordingly, the test was expected aqueous control solutions each containing 5 mg Cl L-1 were to reflect the universal applicability of the proposed volatility prepared from diVerent organochlorine compounds in 0.2% correction procedure.m/m propan-2-ol in water. Owing to the low water solubility The results for synthetic samples summarized in Table 2 of these compounds, stock standard solutions were first pre- were obtained with and without applying the volatility correcpared in propan-2-ol. They were then serially diluted to tion procedure. These samples were measured with reference required concentrations. The final solutions contained no more to a single calibration solution prepared from a non-volatile than 2000 mg L-1 of propan-2-ol and therefore water could silicon or chlorine compound.For each sample a correction be regarded as the solvent. Control solution 1 was prepared curve was prepared with silicon or chlorine compounds that with 1,1,1-trichloroethane (bp 75 °C), solution 2 with chloro- diVer in volatility from the compounds used to prepare the form (bp 62 °C), solution 3 with dichloromethane (bp 40 °C) calibration solution.and solution 4 with carbon tetrachloride (bp 77 °C). A cali- The results obtained validate the proposed scheme for bration solution containing 5.0 mg Cl L-1 was prepared from handling samples with widely diVerent volatilities or boilingammonium chloride in 0.2% m/m propan-2-ol in water. The points. The uncorrected Si and Cl concentration results in blank was 0.2% m/m propan-2-ol in de-ionized water. Table 2 exhibit an average positive error of almost 800%. For Six synthetic sample solutions listed in Tables 2 and 3 were each solvent series the error for uncorrected concentrations prepared in 0.2% m/m propan-2-ol in water.Samples 4 and 5 increased as the compound boiling-point decreased. This error contained a mixture of 3.4 mg Cl L-1 as NH4Cl and 0.097 mg is largest for compounds in m-xylene (average error 1660%, Cl L-1 as dichloromethane, and 3.28 mg Cl L-1 as NH4Cl range 310–3000%), followed by water (570% for Cl ), dec-l-ene and 0.118 mg Cl L-1 as chloroform, respectively.(335%) and hexan-1-ol (315%). The corresponding error for A synthetic sulfur sample solution (No. 6 in Table 3) conthe corrected Si and Cl concentrations ranged from -55 to taining 2.36 mg S L-1 was prepared from CS2 (bp 40 °C) to 120% in the four solvents. The largest errors are observed for test the validity of the correction factor obtained from analysis compounds in dec-1-ene (average error 80%, range 25–120%) of another element (i.e., chlorine) and compared with a stanand the smallest in water (average error -3%, range 12 to dard solution containing 9.9 mg S L-1 as (NH4)2SO4.-41%). Overall the average error for corrected values was 15% In xylene, water, and hexan-1-ol, the corrected values Method tended to be more overcorrected (biased low) with increasing The instruments are stabilized for drift by allowing them to compound boiling-point. In dec-1-ene the corrected values run for at least 1 h before the samples, calibration solution, were generally biased high regardless of the sample compound blank and control solution are measured.The volatility correc- boiling-point. Thus, as the analyte volatility decreases, the tion procedure is then applied. The control, sample and correction generally becomes less accurate. However, the corstandard solutions are measured at two diVerent spray- rection is significantly more accurate than using the uncorchamber temperatures, T1=0 °C and T2=room temperature rected values.Since volatility correction is less critical for low- (22 °C). The control solution measurements are used to calcu- than high-volatility analytes, this tendency is not regarded as late DI/I1 [or (DI/I1)x100] and I1/I10 for each control solution. a serious drawback. The large diVerences between the corrected I10 is the measured intensity of the calibration solution at and uncorrected data indicate the usefulness of the proposed temperature T1.The calculated control solution I1/I10 values scheme for volatility eVect correction. are plotted against the DI/I1 [or (DI/I1)x100] values to con- Application of the correction scheme to adjust for the struct the linear regression function. Next the relative change volatility eVect of a mixture of several compounds containing (DI/I1) in sample intensity I1 is measured. This value is then the same analyte with diVerent volatilities was discussed in the used to obtain the correction factor I10/I1 from the constructed Theoretical section. This situation is frequently encountered correction curve.The correction factor I10/I1 is multiplied by with practical, complex samples. The hypothesis was tested by the measured sample intensity I1 to transform it to I10. This applying the correction approach and eqn. (17) to five analyte correction transforms I1 to a value expected for a non-volatile samples prepared from a mixture of two compounds with compound.The corrected intensity I10 is then used to calculate diVerent volatilities. Three mixtures of organosilicon com- the correct concentration by reference to the non-volatile pounds and two mixtures of organic and inorganic chlorine standard. compounds were analyzed for Si and Cl, respectively. The In the measurement of practical samples, typically the results are summarized in Table 3. The error in the Si determi- analyte compounds in the samples are unknown and probably nation was 192% on average without correction, but -27% not among the compounds used to construct the correction with correction.The uncorrected errors were approximately curve. Accordingly, to simulate these sample types, the correcthe same as for the less volatile single Si compounds of the tion factors for synthetic samples analyzed were determined mixture given in Table 2, whereas the corrected Si concen- from correction curves but excluding the compound in the trations were smaller than expected. The error in the Cl sample.Thus, Fig. 1–4 with one fewer datum were used for determination was on average 22% without correction, far less the synthetic sample analysis. than expected for the volatile organochlorine compounds alone (Table 2) and-22% with correction. The corrected Cl concen- Results and discussion trations also were lower than expected from the corrected values in Table 2. These results indicate the successful feasibil- The universal applicability of the volatility correction proity of applying the volatility correction to analyte mixtures in cedure was tested by measuring samples prepared from analyte diVerent chemical forms.However, the unexpected generally compounds with diVerent volatilities and chemical structures using solvents with diVerent physical and chemical properties. low corrected concentrations require further investigation. J. Anal. At. Spectrom., 1999, 14, 793–799 797Table 2 Silicon and chlorine concentrations before and after volatility correction for pure compounds in diVerent solvents Expected Uncorrected Corrected concentration/ concentration/ Error concentration/ Error RSD (%) No.Simulated sample Bp/°C Solvent Analyte mg L-1 mg L-1 (%) mg L-1 (%) (n=3) 1 Bis(trimethylsilyl )methane 132 m-Xylene Si 20.8 86.0 310 9.4 -54.8 2.8 2 Dichlorodimethylsilane 70.5 m-Xylene Si 4.2 68.0 1520 3.1 -26.2 1.9 3 Chlorotrimethylsilane 58 m-Xylene Si 5.8 109.0 1780 5.0 -13.8 2 4 Tetramethylsilane 26 m-Xylene Si 4.2 132.0 3040 4.9 16.7 1 1 Hexamethyldisilane 112 Dec-1-ene Si 23.6 56.3 139 51.1 117 1.6 2 Trimethoxymethylsilane 103 Dec-1-ene Si 32.8 74.0 126 41.1 25.3 1.8 3 Tetramethylsilane 26 Dec-1-ene Si 7.0 58.9 740 13.9 98.6 2.1 1 Tetraethylsilane 153 Hexan-1-ol Si 20.6 25.5 23.8 16.5 -19.9 2.9 2 Hexamethyldisilane 112 Hexan-1-ol Si 14.9 71.6 380 17.0 14.1 2.1 3 Trimethoxymethylsilane 103 Hexan-1-ol Si 21.1 63.4 200 31.7 50.2 1.8 4 Tetramethylsilane 26 Hexan-1-ol Si 5.7 43.0 654 5.7 0.0 1.6 1 1,1,1-Trichloroethane 75 Water Cl 4.2 13.7 226 2.5 -40.5 2.7 2 Chloroform 62 Water Cl 5.1 37.1 628 5.7 11.8 2.2 3 Dichloromethane 40 Water Cl 4.9 46.3 844 5.9 20.4 1.4 Table 3 Element concentrations before and after volatility correction for mixtures of analyte silicon and chlorine compounds and sulfur in carbon disulfide Expected Found/mg L-1 Error (%) concentration/ RSD (%) No.Simulated sample mixture Solvent Analyte mg L-1 Uncorrected Corrected Uncorrected Corrected (n=3) 4 Trimethoxymethylsilane+ Dec-1-ene Si 23.8 68.6 16.8 188 -29.4 1.4 tetramethylsilane 5 Hexamethyldisilane+ Dec-1-ene Si 18.2 56.1 14.7 208 -19.2 1.8 tetramethylsilane 5 Trimethoxymethylsilane+ Hexan-1-ol Si 16.6 46.6 11.3 181 -31.6 2.3 tetramethylsilane Average: 192 -26.9 4 Dichloromethane+ Water Cl 3.5 4.3 2.3 22.9 -34.3 2.8 ammonium chloride 5 Chloroform+ammonium Water Cl 3.4 4.1 3.1 20.6 -8.8 3.1 chloride Average: 21.7 -21.6 6 Carbon disulfide Water S 2.36 8.2 2.1 248 -11.0 2 A correction curve derived from the control solutions of m-xylene.The accuracy of the analysis cannot be checked one element theoretically can be applied to correct for the without suitable reference materials. Nevertheless, the results volatility of other elements. This hypothesis was tested were compared with those obtained from the same sample experimentally by determining a volatile sulfur compound hydrolysis using dilute nitric acid followed by evaporating the (CS2 in water) using a correction curve derived from the solution to near dryness and dissolving the residue in 5% nitric chlorine compounds control solutions.The results (Table 3, acid. The comparison indicates that the corrected results sample 6) show that the error in the uncorrected sulfur compare reasonably well (within an order of magnitude) to concentration decreased from 240 to -11% for the corrected the values from the hydrolysis procedure.concentration despite the diVerence in chlorine and sulfur Some diVerences between the expected and the corrected ICP-MS sensitivities. values in Tables 2 and 3 are greater than statistical imprecision The correction approach also was tested by applying it to of the correction factor regression. Fortunately, this error is analyze practical samples exhibiting analyte volatility. A less pronounced for highly volatile compounds where the sample of semiconductor-grade trimethylindium was diluted correction is practically needed compared with mediumto 10% m/m in xylene and analyzed for trace impurities by volatility compounds.ICP-AES. This sample is believed to contain volatile Al, Zn, Preparation of correction curves with reactive organometal- Si and Ca impurities based on earlier experiments.12 The lic compounds can be diYcult and might contribute bias. results, summarized in Table 4, were corrected using a correc- Some of the silicon compounds (e.g., chlorinated organosiltion curve derived from silicon compound control solutions in icons, trimethoxymethylsilane) used in this feasibility study react with residual trace water in the anhydrous organic Table 4 Corrected ICP-AES analysis of trimethylindium solvents (e.g., m-xylene).Therefore, the m-xylene was pretreated with octadecyltrichlorosilane to remove water before Concentration control solutions were prepared. Accordingly, a small Si found/mg g-1 Concentration blank was subtracted from each of the control solution Analyte after hydrolysis/ readings.The variability of this blank and the original wavelength/nm Uncorrected Corrected mg g-1 residual water in anhydrous solvents probably contributed to Zn I 213.856 3.2 0.02 <0.02 bias and scatter in Fig. 1–4. Alternative, unreactive organosil- Si I 251.612 0.9 0.06 <0.25 icon compounds are being evaluated to eliminate hydrolysis Al I 309.283 36 0.29 1.2 with residual water in anhydrous solvents. Furthermore, Al I 396.152 39 0.25 2.8 organochlorine compounds in water have a tendency to 798 J.Anal. At. Spectrom., 1999, 14, 793–799adhere to the spray chamber walls and result in a Cl memory unreactive compounds. Both of these aspects are under investigation. eVect. Evaluation of these potential sources of bias requires further study. Experimental feasibility has been demonstrated for chlorine, silicon and sulfur compounds in synthetic aqueous and organic solutions, and Al, Si and Zn in a reactive organometallic Conclusions compound.Tests are under way to extend the practical application of this correction approach to other elements and The error from the diVerence in volatilities between the analyte solvents. Furthermore, diVerences in correction accuracy with compound in the sample and the calibration solution can be solvents and mixtures observed also require additional study. corrected using a scheme that is based on measurements at two diVerent spray chamber temperatures.The proposed Acknowledgement method is simple to apply, fast and can be adopted for routine analysis. The authors thank Perkin-Elmer Corporation (Norwalk, CT, The volatility correction approach is reasonably accurate USA) for providing the Optima 3000 prototype system. The and general. Overall the average error was 15% for corrected authors also acknowledge Spectro Analytical Instruments values compared with 750% for uncorrected concentrations.(Fitchburg, MA, USA) for providing the Spectromass 2000 The range in concentration errors was also significantly smaller ICP-MS system. This investigation was supported by ICP than for uncorrected concentrations. The approach allows for Information Newsletter, Inc. the correction of a mixture of several compounds containing the same analyte with diVerent volatility as indicated by eqn. References (17). Again, the concentration error is generally reduced for corrected results. 1 I. Bertenyi and R. M. Barnes, Anal. Chem., 1986, 58, 1734. Also, a correction curve derived from the control solutions 2 K. Takeda, M. Minobe, T. Hoshika, T. Jinno and T. Yako, Analyst, 1990, 115, 535. of one element (e.g., Cl ) can be applied to correct for the 3 J. McLean, H. Zhang and A. Montaser, Anal. Chem., 1998, 70, volatilities of other elements (e.g., S). This can be inferred 1012. from the observation that eqns. (14) and (17) contain variables 4 K. LaFreniere, V. A. Fassel and D. E. Eckels, Anal. Chem., 1987, and constants that depend only on the volatility property of 59, 879. the compound and are independent of the chemical composi- 5 J. Christodoulou, M. Kashani, B. Keohane and P. Sadler, J. Anal. tion, identity of the analyte and spectral line or the isotope At. Spectrom., 1996, 11, 1031. 6 T. W. Avery, C. Chakrabarty and J. J. Thompson, Appl. used for measurement. The dependence of the sensitivity of Spectrosc., 1990, 44, 1690. the ICP-AES and ICP-MS measurement on the nature of the 7 CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC analyte and the spectral line/isotope used is eliminated by the Press, Boca Raton, FL, 73rd edn., 1992–93, Table 6.8b. use of relative units for correction factor, I10/I1, and the 8 A. Boorn and R. Browner, Anal. Chem., 1982, 54, 1402. change in intensity, DI/I1. This capability of applying a single 9 A. Cresser and R. Browner, Spectrochim. Acta, Part B, 1980, correction curve derived from control solutions of one element 35, 73. 10 A. Boorn, A. Cresser and R. Browner, Spectrochim. Acta, Part B, to correct for the volatility of other elements makes the 1980, 35, 823. proposed scheme very practical. 11 M. D. Argentine, A. Krushevska and R. M. Barnes, J. Anal. At. The correction curve need be determined only once. Spectrom., 1994, 9, 112. Thereafter, the same curve could be used repeatedly provided 12 M. D. Argentine and R. M. Barnes, J. Anal. At. Spectrom., 1994, that the experimental measurement conditions and the solvent 9, 1371–1378. type used remain unchanged. Correction curves can be improved, however, by using more calibration compounds and Paper 8/08391F J. Anal. At. Spectrom., 1999, 14, 793–799 799

ISSN:0267-9477    

DOI:10.1039/a808391f

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Universal calibration for analysis of organic solutions of medium and low volatility by inductively coupled plasma-atomic emission spectrometry† Assad A. Al-Ammar, Rajesh K. Gupta and Ramon M. Barnes* Department of Chemistry, Lederle Graduate Research Center Towers, University of Massachusetts, Box 34510, Amherst, MA 01003–4510, USA Received 29th October 1998, Accepted 26th February 1999 A novel chemometric technique is described to facilitate the use of a single organic solvent matrix standard to calibrate inductively coupled plasma atomic emission spectrometry (ICP-AES) for the accurate determination of trace elements in another organic solvent or complex mixture of several organic solvents. Analysis errors arising from the diVerence in solvent matrix vapor loading and chemical composition between standard and sample are corrected.The technique is intended for organic solvent mixtures for which the plasma vapor loading is tolerable with conventional sample introduction techniques (i.e., conventional nebulizer–spray chamber arrangement without a solvent desolvator).Simultaneous measurement of atomic and ionic spectral lines of the same analyte is required. A correction factor is estimated from its linear correlation with the line intensity ratio. For elements without a sensitive atom line, the correction is estimated from the correlation between their ionic line intensities and the ratio of ionic-to-atomic line intensities of another analyte in the same sample.Experimental tests with seven trace elements (Al, Be, Ca, Cu, Fe, Mg, and Mn) in diVerent organic solvent mixtures (xylene, dichloromethane, hexane, carbon disulfide, acetone and 1,2,3,4-tetrahydronaphthalene) demonstrate the eVectiveness of universal calibration. The determination of trace elements in hydrocarbons is import- ated with DIN. Furthermore, a major disadvantage of DIN is the plasma cooling that occurs when organic solutions are ant in the petroleum and petrochemical industry.Elements that are aggressive catalysts, poisons or corrosive in the nebulized. This degrades the detection limit for most elements by as much as an order of magnitude compared with conven- cracking furnace are harmful even at mg L-1 levels owing to the large volumes of hydrocarbons processed. Also, shipments tional sample introduction methods.6 The aim of this investigation was to demonstrate universal of purchased feed or outgoing products frequently require monitoring of trace contaminants.However, the determination calibration with a conventional nebulizer–spray chamber without desolvation. The nebulizer–spray chamber arrangement, of trace elements in refinery and chemical plant streams is complicated by the continuous variability in composition and in contrast to the use of a desolvator, permits the analysis of volatile analyte species. However, the organic liquids that can volatility, which makes the use of a single organic solvent matrix standard for universal calibration impractical.be analyzed are limited to those that can be tolerated by the plasma (i.e., organic compounds with medium or low vola- Universal calibration is defined as using standards prepared in a solvent to measure samples in diVerent solvent matrices tility). Nevertheless, these compounds constitute a large class encountered in the petroleum and petrochemical industry. than the calibration standard.Various approaches have been applied to achieve universal Furthermore, petroleum compound mixtures with medium or low volatility containing small and varying percentages of calibration. Botto and Zhu1,2 used a membrane desolvator to strip organic solvent from the aerosol with an eYciency of highly volatile compounds also can be analyzed. A chemometric technique was developed to correct for the 99.9%. Cryogenic desolvation at temperatures below -80 °C also was used to remove organic vapor.3 Since these measurement errors arising from the diVerence in matrix loading, chemical composition and nebulization eYciency approaches are based on solvent stripping, volatile analyte compounds, which usually form the largest fraction of the between the sample and standard matrices.Universal calibration can be achieved. total analyte concentration, also are lost. Consequently, solvent removal methods are of limited practical use. Botto and Zhu1 also evaluated the influence of chemical composition on solvent matrix loading.They discovered that aromatic com- Theory and derivation pounds suppressed analyte signals more than aliphatic com- DiVerences between sample and calibration solution matrices pounds. Chlorinated and oxygenated organic compounds can cause variations in transport eYciencies and plasma matrix suppressed analyte signals less than simple hydrocarbons of loading. A large error can be expected in estimating the analyte comparable volatilities.Direct injection nebulization (DIN) concentrations. Accordingly, to make universal calibration also can be applied for universal calibration, because it eliminpossible a correction must be estimated for each analyte in ates the variation in matrix loading that originates from the every sample matrix. The analyte signal measured for a sample diVerence in volatility.4,5 However, matrix loading that results must be multiplied by a corresponding correction factor to from the diVerence in chemical composition cannot be eliminconvert it to the value expected if the sample matrix were the same as the standard. The corrected signal can then be used †Presented in part at the Fifth Rio Symposium on Atomic to obtain the correct concentration by reference to the cali- Spectrometry, Cancu� n, Mexico, October 4–10, 1998, and the 1999 bration standard. Thus, the correction factor eliminates the European Winter Conference on Plasma Spectrochemistry, Pau, France, January 10–15, 1999.diVerence in matrix loading and transport eYciency between J. Anal. At. Spectrom., 1999, 14, 801–807 801sample and standard. The aim of this work was to develop a correction factor with the measured spectral lines intensities. This approach was used previously to develop a chemometric chemometric method to establish a mathematical relationship that links the correction factor to matrix loading and transport technique called common analyte internal standardization (CAIS).It was used to correct for the signal drift9 and eYciency. This relationship must be expressed as variables that can be measured, for example, by inductively coupled inorganic matrix eVect (NaCl, H2SO4, HNO3, etc.) in ICPAES. 8 plasma atomic emission spectrometry (ICP-AES). Since only spectral line intensities are measured in AES, this dependence Vapor loading should use spectral line intensities as variables. The diVerence in transport eYciency has been corrected From the diVerence in the magnitude of intensity depression with an internal reference element.1 For this purpose, one (or enhancement) between an atomic and ionic line of the internal reference element is suYcient to correct for all the same element, a mathematical function that relates vapor analytes in a sample.2 The transport eYciency can be corrected loading to a correction factor can be developed starting from by multiplying the analyte intensity by the ratio of intensities eqn.(1). of the internal reference in the standard and the sample. The For an ionic line correction should be applied after correcting the intensities of the internal reference and the analyte for matrix loading. The Ii=Ii0 exp(Ai-k1iC1-k2iC2-…-kniCn) (2) problem that remains to be solved in this study is to correct and for an atomic line for matrix loading by finding an algebraic relationship between matrix loading and a correction factor.Ia=Ia0 exp(Aa-k1aC1-k2aC2-…-knaCn) (3) Dividing eqn. (2) by eqn. (3) and rearranging terms yields Organic matrix loading (Ia0/Ii0)(Ii/Ia)=exp[Ai-Aa+C1(k1a-k1i)+…+Cn(kn a-kni)] To find the correction function, a mathematical analysis should (4) be based on an exact algebraic representation of the analyte intensity, I, and the matrix concentration, C. The eVect of From eqn. (2) inorganic matrices (such as, NaCl, Cl2) on analyte intensity Ii/Ii0=exp(Ai-k1iC1-k2iC2-…-kniCn) (5) in ICP-AES resembles an exponential decay curve.7 Correspondingly, we postulate that organic matrix vapor After subtracting eqn.(4) from eqn. (5), Ii0/Ii can be expressed loading also depresses the analyte intensity according to an as exponential decay curve given by the equation Iio/Ii=(Ia0/Ii0)(Ii/Ia)+exp[-(Ai-k1iC1-k2iC2-…-kniCn)] I=I 0 exp(A-k1C1-k2C2-...-knCn) (1) -exp[Ai-Aa+C1(k1a-k1i)+…+Cn(kna-kni)] (6) Eqn. (1) describes the situation where the organic matrix of Eqn.(6) is useful because it indicates that the correction the universal calibration solution diVers from the organic factor Ii0/Ii can be calculated from the experimentally deter- matrix of the sample. The sample matrix (major matrix) may mined ratio Ii/Ia. Another helpful version of eqn. (6) can be exist in pure form or mixed with small concentrations, C1, obtained by dividing eqn. (3) by eqn. (2) followed by sub- C2, …, Cn, of other organic solvents (minor matrices) that are tracting the resulting equation from eqn.(5) and rearranging: more volatile than the sample matrix. An example is premium grade gasoline containing approximately 1% of pentanes and Ii0/Ii=(Ii0/Ia0)(Ia/Ii)+exp[-(Ai-k1iC1-k2iC2-…-kniCn)] butanes. In eqn. (1), I0 represents the intensity of the analyte -exp[Aa-Ai+C1(k1i-k1a)+…+Cn(kni-kna)] (7) dissolved in the organic matrix of the universal calibration solution. I represents the intensity of the same concentration The ratio Ii/Ia [eqn.(6)] or its inverse, Ia/Ii [eqn. (7)] is not of the analyte dissolved in sample matrix (major matrix), aVected by the change in transport eYciency, since atomic and which contains concentrations C1, C2, …, Cn of other volatile ionic lines are aVected by transport eYciency to the same organic compounds. If C1, C2, …, Cn are zero, then I= extent. Therefore, these ratios are true specific measures of I0 expA represents the intensity of the analyte dissolved in matrix vapor loading.pure sample matrix. A and k1, k2, …, kn are constants related Although eqns. (6) and (7) are similar, the results of a to the organic compounds in the sample. numerical simulation and experiments indicate that these two The validity of eqn. (1) was tested experimentally. The equations behave diVerently. The plots of correction factor eVects of small additions of CH2Cl2 to a m-xylene matrix and Ii0/Ii against Ii/Ia [eqn. (6)] or Ia/Ii [eqn. (7)] give straight lines.hexane to a 1,2,3,4-tetrahydronaphthalene matrix were meas- However, the linear regression correlation coeYcient was ured for several atomic and ionic lines of Be, Ca, Cu, Fe, Mg higher with eqn. (7) than with eqn. (6). Accordingly, the plot and Mn. Small hexane or CH2Cl2 concentration additions of the correction factor against Ia/Ii is preferred to correct for (0–6%) do not aVect the transport eYciency of the matrix. matrix vapor loading. However, a marked increase in vapor loading is expected, because of the high volatility of the additives compared with Application the xylene or tetrahydronaphthalene matrices.The measure- Analysis method. An analytical scheme was developed to ments were followed by evaluating the applicability of the correct for the matrix vapor loading, so that universal cali- exponential decay of eqn. (1). Straight lines were obtained for bration can be applied. The approach employs several control all additives studied with correlation coeYcients 0.98, in solutions and one universal calibration solution.The control excellent agreement with the logarithmic form of eqn. (1): solutions and the calibration solution are prepared from the ln I=ln I 0+A-k1C1-k2C2-…-knCn (1¾) same organic matrix and contain exactly the same concentrations of analytes. In contrast to the calibration solution, Furthermore, the magnitude of signal suppression was found to be related to the ionization and excitation potential as the control solutions contain, in addition to the major matrix, varying percentages of another organic minor matrix that has reported earlier for an inorganic matrix eVect.7,8 Atomic lines are less depressed than ionic lines.This diVerence in behavior higher volatility compared to the major matrix (e.g., hexane, chloroform). Variable amounts of the minor matrix are added between the atomic and ionic lines toward organic matrix vapor loading is used as a key factor in the next section to to the control solutions to create changeable high vapor loadings.However, the amounts of the minor matrix added develop an exact algebraic relationship applied to estimate the 802 J. Anal. At. Spectrom., 1999, 14, 801–807should be small, so that the transport eYciency is unchanged atomic-to-ionic line combination can be used as internal reference. When no fitting analyte can be found in the sample, for all the control solutions and equal to that of the calibration solution. Therefore, the eVect of only vapor loading is meas- an appropriate amount of the internal reference element is added to the sample.However, in contrast to the conventional ured and corrected. The control solutions, calibration solution, samples and method of internal reference addition, the amount added need not be exact, because only the ratio Ia/Ii needs to be measured. blank are measured under the same experimental conditions. From the measurement results for the control solutions and This ratio is independent of the amount added.Also, in contrast to the conventional method of internal reference the calibration solution, a straight-line correction curve is constructed by plotting the vapor loading correction factor, addition, only one internal reference is needed for all the elements in the sample. Ii0/Ii, against Ia/Ii of eqn. (7). Ii0 is the intensity of the analyte in the calibration solution as defined by eqn. (1).The measured value of Ia/Ii for the sample is then used to estimate the vapor Experimental loading correction factor for the sample with aid of the constructed correction curve. The sample correction coeYcient Instrumentation is then multiplied by the measured sample intensity, Ii, to A commercial ICP-AES system (Optima 3000, Perkin-Elmer, transform it to Ii0, which corresponds to the predicted intensity Norwalk, CT, USA) was used for all the experiments.The of the analyte in the sample if it exists in the matrix of the experimental operating parameters (Table 1) were selected calibration solution. The corrected intensity is then used to based on the optimization reported in the manufacturer’s calculate the vapor loading-corrected concentration by referinstrument manual. Experimental wavelengths are listed in ence to the calibration solution. This concentration is further Table 2. corrected for transport eYciency, if the sample and the calibration solution transport eYciencies are diVerent from each Table 1 Operating parameters for the ICP-AES measurements other, by using an internal reference as mentioned at the beginning of the section.ICP system Optima 3000 prototype Rf power 1.3 kW Frequency (free running) 40 MHz Analyte as internal reference. This scheme applies to the ICP torch Type 2 quartz slotted extension analytes with a pair of sensitive atomic and ionic lines, which Torch injector Ceramic alumina is true for most elements.However, for some elements either Outer argon flow rate 15 L min-1 only an atomic line or ionic line can be found with appropriate Intermediate argon flow 2.0 L min-1 sensitivity. In this situation obtaining the correction curve for rate the analyte is possible by using the ratio Ia/Ii, of an other Observation height 15 mm Central argon flow rate 0.8 L min-1 analyte in the sample. This approach, which specifies an Nebulizer Concentric glass (Glass Expansion, analyte to act as a surrogate internal reference, seems to be Hawthorn, Victoria, Australia, reasonable, because the ratio Ia/Ii is a general measure for Model 38493) vapor loading no matter to which analyte it belongs.Sample pump rate 0.8 mL min-1 To apply this concept for an analyte with only ionic lines Sample pump tubing Viton, orange–orange (0.035 in id) of suitable intensity, a mathematical correlation should first Spray chamber Glass Scott double pass coolant jacketed (Spectro Analytical Instruments, be found to relate the analyte correction factor, Iiom/Iim to the Fitchburg, MA, USA) internal reference correction factor, Iior/Iir. This correlation Spray chamber temperature 0 °C can be obtained from the equation Integration time (auto) 10 s Background correction ±0.04 nm (Iior-Iir)/Iir=(Iiom-Iim)/Iim (8) Drain Pumped Eqn.(8) is the equation for the conventional method of internal standardization in ICP-AES. As recognized by several Table 2 Analyte and emission wavelengths investigators,10,11 eqn.(8) lacks general applicability. In this mathematical derivation a more general form is Element Wavelength/nma (Iior-Iir)/Iir=H(Iiom-Iim)/Iim (9) Al I 309.271 where H is a constant the magnitude of which depends on the Be I 234.861, II 313.042 Ca II 393.366, II 396.847 nature of the analyte and the internal reference. Cu I 324.754, I 327.396 From eqns. (7) and (9), the equation Fe II 259.940 Iiom/Iim={(Iior/Iaor)(Iar/Iir)+exp[-(Air-k1irC1r-k2irC2r Mg II 279.553, I 285.213 Mn II 257.610 -…-knirCnr)] -exp[Aar-Air+C1r(k1ir-k1ar)+… (10) aI are atom lines and II are ion lines.+Cnr(knir-knar)]}-1+H] can be derived. Numerical simulations based on eqn. (10) Table 3 EVect of 2% m/v of various volatile compounds in a xylene indicate that the plots of the correction factor for the matrix on the analyte line signal intensity analyte, m, against the ratio, Iar/Iir, for the internal reference Signal change (%) element, r, provide straight lines.Hence the same correction scheme based on eqn. (7) can be applied. Analyte and CH2Cl2 Hexane CS2 Acetone If the analyte has only atomic lines of proper sensitivity, an wavelength/nm (bp 40 °C) (bp 69 °C) (bp 46 °C) (bp 56 °C) equation similar to eqn. (10) can be derived that correlates the analyte correction coeYcient measured by using atomic Be I 234.861 -17 -19 +23 -2 Be II 313.042 -36 -45 +13 -19 lines with the atomic-to-ionic line ratio of the internal refer- Mg I 285.213 -4 +3 +44 +11 ence element.Mg II 279.553 -50 -51 +24 -21 In the modified correction scheme based on eqn. (10), Fe II 259.940 -42 -46 +35 -18 adding an internal reference element to the sample is not Ca II 393.366 -43 -41 +42 -13 necessary. Any other analyte in the samples with suitable J. Anal. At. Spectrom., 1999, 14, 801–807 803Table 4 Experimental values of A in xylene and k in 2% m/v of various volatile compounds in xylene k Analyte and wavelength/nm A CH2Cl2 Hexane CS2 Acetone Be I 234.861 -0.39 0.099 0.11 -0.108 0.01 Be II 313.042 -0.94 0.23 0.30 -0.048 0.12 Mg I 285.213 -0.05 0.011 -0.011 -0.19 -0.054 Mg II 279.553 -0.92 0.35 0.36 -0.10 0.124 Fe II 259.940 -0.91 0.28 0.31 -0.144 0.10 Ca II 393.366 -0.66 0.28 0.26 -0.17 0.081 Simulation values— Atom Aa=-0.387 ka=0.04 Ion Ai=-0.94 ki=0.1 Fig. 1 Log(signal intensity) for the Mg 279 nm line as a function of Fig. 3 Numerical simulation of correction factor, Iio/Ii, as a function percentage of dichloromethane.of Ia/Ii. Fig. 2 Log(signal intensity) for the Al 309 nm line as a function of percentage of hexane. Fig. 4 Correction factor Ii0/Ii for Be as a function of Ia/Ii for Be. The six points correspond to Be in pure xylene and 2% hexane, 4% hexane, 2% CS2, 2% acetone and 2% CH2Cl2, in xylene. Error bars=2%. Reagents Commercial atomic emission calibration solutions prepared from a non-volatile metal organic salt in hydrocarbon oil act as the volatile minor matrix.Three control solutions were prepared containing 4, 8 and 12% m/m hexane. Several samples (VHG Laboratories, Manchester, NH; 100 mg L-1) were used to prepare three types of test solutions: sample, calibration containing 4 or 6 mg L-1 of each element were prepared in m-xylene (bp 139 °C, 99% purity; Aldrich) as a major matrix and control. Each of the control solutions contained a mixture of Be, Mg, Ca, Fe,Mn, Cu and Al.The control and calibration containing 2% m/m of various highly volatile organic compounds to act as the minor volatile matrix. These volatile solutions contained 4.0 mg L-1 of each element. The calibration solution was prepared in a 1,2,3,4-tetra- minor matrices were CS2 (bp 46 °C), hexane, acetone (bp 56 °C) and CH2Cl2 (bp 40 °C). hydronaphthalene matrix (tetralin, bp 217 °C, 98% purity; Aldrich, Milwaukee, WI, USA). The control solutions were To match the diVerence in transport eYciency of the calibration solution and the sample, each sample and the cali- prepared in 1,2,3,4-tetrahydronaphthalene as the major matrix to which hexane (bp 69 °C) was added in various amounts to bration solution contained 4 mg L-1 indium as an internal 804 J.Anal. At. Spectrom., 1999, 14, 801–807Verification The exponential decay of eqn. (1) was tested under the experimental and instrumental arrangements given in Table 1 that are typical of those widely used for routine analysis.The eVect of 0–5.8% CH2Cl2 in m-xylene and 0–12% hexane in 1,2,3,4-tetrahydronaphthalene was measured for several ionic and atomic lines and results plotted as ln I against the CH2Cl2 or hexane additive. Numerical simulation Although eqns. (6) and (7) closely resemble each other, their behaviors were tested by numerical simulation under conditions that simulate typical operations used in routine analysis. In this numerical test the sample was taken to be an organic matrix that diVers in volatility from that of the calibration solution.The sample matrix was assumed to contain various concentrations (from 0 to 5%) of a minor organic Fig. 5 Correction factor Ii0/Ii for Mg as a function of Ia/Ii for Mg. matrix that has a volatility much higher than that of the major The six points correspond to pure xylene and 2% hexane, 4% hexane, matrix. In this simulation, constant A and k values were 2% CS2, 2% acetone and 2% CH2Cl2 in xylene.Error bars=2%. experimentally determined (Table 3) when the exponential decay behavior of the organic matrix was examined (Ai, Aa, ka and ki in Table 4). Also, the ionic spectral lines were assumed to be more aVected by the matrix than atomic lines corresponding to experimental observations. Results and discussion The method was tested by measuring trace concentrations of Al, Be, Ca, Cu, Fe, Mg and Mn in simulated samples prepared from organic matrices that diVer in chemical composition, volatility, surface tension and viscosity from the matrix of the calibration solution.Also, several highly volatile organic compounds were mixed with the major sample matrix to simulate the composition of petroleum compounds that include medium and low volatility products containing small but variable percentages of highly volatile organic compounds (e.g., pentanes and butanes). The presence of a variable composition of highly volatile organic compounds causes changeable vapor loadings and a universal standardization strategy is therefore Fig. 6 Correction factor Ii0/Ii for Mg as a function of Ia/Ii for necessary. The highly volatile organic compounds used in this Mg. The four points correspond to Mg in pure 1,2,3,4- study diVered not only in volatility but also in chemical tetrahydronaphthalene and in 4, 8 and 12% hexane in 1,2,3,4- composition, thus testing the ability of the proposed method tetrahydronaphthalene. Error bars=2%.to correct for both vapor loading and chemical composition. Exponential response reference. Pure xylene and 1,2,3,4-tetrahydronaphthalene were measured as blanks. The eVect of highly volatile compound additions on analyte The eVect of 0–5.8% CH2Cl2 in m-xylene matrix was meas- (Be, Fe, Mg, Ca) signals is reported in Table 3 for four ured for several ionic and atomic lines of 4 mg L-1 Be, Ca, additives to a xylene matrix. The suppression (or enhancement) Cu, Fe, Mg and Mn (Table 2) to test eqn.(1). The experiment is less for atomic than ionic lines. The applicability of the was repeated using 0–12% hexane in a 1,2,3,4-tetraexponential decay of eqn. (1) was evaluated by plotting ln I hydronaphthalene matrix. against the percentage of CH2Cl2 or hexane additive. Examples selected from the straight line plots obtained are given in Method Fig. 1 for Mg with CH2Cl2 and Fig. 2 for Al with hexane in xylene. From these plots, values of A and k were calculated, The instrument was first stabilized for drift by allowing it to run for 1 h, after which the samples, calibration solutions, as illustrated for xylene in Table 4.The A and k values of Be, Ca, Fe and Mg ion lines are similar but significantly diVerent blanks and control solutions were measured. The procedure for matrix correction was then followed in exactly the way from those of Be and Mg atom lines. The A and k values for atomic and ionic lines (Ai, Aa, ka, and ki Table 4) used for discussed in the Theory and derivation section.The elements measured were subdivided into three groups simulations were estimated from these data. The magnitude of signal suppression (or enhancement) is according to the way in which the correction experiment was conducted. The first group, Mg and Be, was corrected by related to the ionization and excitation potential as reported for an inorganic matrix eVect.7,8 Atomic lines are less depressed using their own atomic/ionic line intensities.The second (Fe, Mn and Ca) and third (Al and Cu) groups were corrected by than ionic lines. This diVerence in behavior between atomic and ionic lines toward the organic matrix vapor loading eVect using the ratio Ia/Ii of Mg (i.e., Mg was regarded as the internal reference) to simulate elements that have no atomic is pivotal in developing the relationship applied to estimate the correction factors for measured spectral line intensities. or ionic lines of proper sensitivity, since the only commercial calibration solutions available for this study contained no These results also indicate the importance of the chemical composition of the matrix on the degree of depression or such elements.J. Anal. At. Spectrom., 1999, 14, 801–807 805Table 5 Trace element concentrations in xylene and in volatile organic–xylene mixtures using standards in a 1,2,3,4-tetrahydronaphthalene matrix Concentration/mg L-1 Standard Mg II 279.553 Be II 313.042 Fe II 259.940 deviation Solution Concentration nm nm nm Mean of mean error Xylene (pure) Expected 6.00 6.00 6.00 Uncorrected 2.50 3.20 3.72 Error (%) -58 -47 -38 -48 10 Corrected 6.00 5.70 6.10 Error (%) 0.0 -5.0 1.7 -1.1 3.5 2% CS2 in xylene Expected 4.00 4.00 4.00 Uncorrected 0.80 0.94 0.93 Error (%) -80 -77 -77 -78 2.0 Corrected 4.20 3.80 4.20 Error (%) 5.0 -5.0 5.0 1.7 5.8 2% acetone in xylene Expected 4.00 4.00 4.00 Uncorrected 0.98 1.15 1.22 Error (%) -76 -71 -70 -72 3.1 Corrected 3.80 4.10 3.70 Error (%) -5.0 2.5 -7.5 -3.3 5.2 2% CH2Cl2 in xylene Expected 4.00 4.00 4.00 Uncorrected 0.98 1.05 1.08 Error (%) -76 -74 -73 -74 1.3 Corrected 4.10 4.30 4.20 Error (%) 2.5 7.5 5.0 5.0 2.5 2% hexane in xylene Expected 4.00 4.00 4.00 Uncorrected 0.84 0.98 0.95 Error (%) -79 -76 -76 -77 1.8 Corrected 3.90 4.00 4.30 Error (%) -2.5 0.0 7.5 1.7 5.2 Ia/Ii ratio (nm/nm) Mg 285/Mg 279 Be 234/Be 313 Mg 285/Mg 279 Number of replicates=3.Standard deviation ranged from 1 to 3%.Table 6 Trace element concentrations in xylene using calibration solutions in a 1,2,3,4-tetrahydronaphthalene matrix Concentration/mg L-1 Concentration/mg L-1 Ia/Ii Element and ratio wavelength/nm Expected Uncorrected Error (%) Corrected Error (%) (nm/nm) Mg II 279.553 4.0 2.0 -50.0 4.4 10.0 Mg 285/Mg 279 Be II 313.042 4.0 2.5 -37.5 4.7 17.5 Be 234/Be 313 Ca II 393.366 4.0 3.5 -12.5 4.0 0.0 Mg 285/Mg 279 Ca II 393.366 6.0 4.6 -23.3 5.6 -6.7 Mg 285/Mg 279 Ca II 396.847 4.0 3.7 -7.5 3.9 -2.5 Mg 285/Mg 279 Ca II 396.847 6.0 4.7 -21.7 5.4 -10.0 Mg 285/Mg 279 Mn II 257.610 6.0 4.0 -33.3 6.1 1.7 Mg 285/Mg 279 Cu I 327.396 4.0 3.6 -10.0 4.9 22.5 Mg 285/Mg 279 Cu I 327.396 6.0 4.5 -25.0 6.2 3.3 Mg 285/Mg 279 Cu I 324.754 4.0 3.6 -10.0 5.0 25.0 Mg 285/Mg 279 Cu I 324.754 6.0 4.5 -25.0 6.5 8.3 Mg 285/Mg 279 Al I 309.283 4.0 2.2 -45.0 4.1 2.5 Mg 285/Mg 279 Al I 309.283 6.0 3.6 -40.0 6.9 15.0 Mg 285/Mg 279 Mean error±SD -26.2±14.1 6.7±10.9 Number of replicates=3.Standard deviation ranged from 1 to 3%. enhancement of the analyte signal. Oxygenated or chlorinated coeYcient=0.76), while the plot of the correction factor against Ia/Ii retains its linear character. Accordingly, it is hydrocarbons cause less signal depression than ordinary hydrocarbons with comparable volatility, in agreement with the preferable to use the plot of the correction factor against Ia/Ii to correct for the matrix vapor loading. results obtained by Boorn et al.12 Numerical simulation Simulated samples The experimental results from the simulated samples were Results of the numerical simulation indicate that plots of the correction factor Ii0/Ii, against Ii/Ia [eqn.(6)] or Ia/Ii [eqn. (7)] used to test the linearity of the relationship between the correction factor and the ratio Ia/Ii. All the constructed curves give straight lines. However, the straight lines obtained from the plot of the correction factor against Ia/Ii [eqn.(7)] (Fig. 3) were linear regardless of the diVerences in volatility and chemical composition. Representative examples of these exper- give better correlation coeYcients than the lines obtained from the plots of the correction factor against Ii/Ia. This diVerence imentally obtained plots are illustrated in Fig. 4–6 for Be and Mg in xylene and 1,2,3,4-tetrahydronaphthalene. This agrees becomes more pronounced when the major and minor solvents of the sample have much higher volatility than the calibration with the results of the numerical simulations with eqn.(7). The simulated samples were measured using a universal solution matrix. Then, the plot of the correction factor against Ii/Ia shows more curvature (curved line with correlation calibration solution prepared in 1,2,3,4-tetrahydro- 806 J. Anal. At. Spectrom., 1999, 14, 801–807naphthalene. The proposed chemometric technique was matrices. The ratio of the atomic-to-ionic line intensities is sensitive to vapor loading and chemical composition of the applied to correct the results for the diVerence in vapor loading and chemical composition between the samples and the stan- matrix.Therefore, the ratio can be used as a mathematical measure for the eVect of these two variables. The chemometric dard matrices. The results of these measurements before and after correction are summarized in Table 5 for Mg, Be and technique based on this observation is capable of facilitating the use of the universal standard calibration strategy on a Fe in xylene and in Table 6 for Mg, Be, Ca, Mn, Cu and Al in xylene.Without correction the signals were suppressed by routine basis. The developed chemometric technique is general, because it can be applied also to elements that have only an 10–80%. Addition of up to 2% of other organic compounds accentuates the errors. Corrected value errors ranged from 0 ionic or an atomic line with suitable sensitivity, but not both.This can be achieved with the ratio of atomic-to-ionic line to 25%, and only for two Cu lines (at 4 mg L-1) were the corrected values worse than the uncorrected values (Table 6). intensities of other analyte in the sample. The new analysis and correction schemes are simple and involve no sample pre- The reported results indicate clearly the eYcacy of this chemometric technique both by using the Ia/Ii, ratio of spectral lines treatment. Hence they are amenable to programmed operation and autosampling.Studies are under way to extend the practi- of the same analyte (Mg and Be) or another analyte in the same sample (Al, Ca, Cu, Fe and Mn). cal application of this chemometric technique to other elements and matrices. The results of applying the proposed corrections to the simulated samples indicate that the transport eYciencies of several matrices (xylene+small percentage of other volatile Acknowledgments compounds) and the standard (1,2,3,4-tetrahydronaphthalene) The authors thank the Perkin-Elmer Corporation for providing are virtually the same despite the marked diVerences in their the Optima 3000 prototype system.This investigation was physical properties. This is expected from the results obtained supported by the ICP information Newsletter, Inc. (Hadley, by Boorn et al.,12 who indicated that even with compounds as MA, USA). diVerent in physical properties as m-xylene (bp 139 °C, viscosity 0.62 cP, density 0.86 g mL-1; and surface tension 29 N m-2) and nitrobenzene (bp 211 °C, viscosity 2.03 cP, References density 1.2 g mL-1 and surface tension 44 N m-2) only a small 1 R. I. Botto and J. Zhu, J. Anal. At. Spectrom., 1994, 9, 905. diVerence exists in their transport eYciencies. Accordingly, the 2 R. I. Botto and J. Zhu, J. Anal. At. Spectrom., 1996, 11, 675. results obtained in this work are concerned with only the 3 D. Wiederin, R. S. Houk, R. Winge and A. D’Silva, Anal. Chem., correction for the diVerence in volatility and chemical composi- 1990, 62, 1155. tion between the sample and the standard matrices. 4 J. Christodoulou, M. Kashani, B. M. Keohane and P. J. Sadler, The eVect of diVerent solvents on the emission background J. Anal. At. Spectrom., 1996, 11, 1031. 5 J. McLean, H. Zhang and A. Montaser, Anal. Chem., 1998, 70, intensity at the wavelengths studied was negligibly small. As 1012. an example, when tetrahydronapthalene solvent was replaced 6 T. Avery, C. Chakrabarty and J. Thompson, Appl. Spectrosc., by xylene, the changes in the background, expressed as equival- 1990, 44, 1690. ent concentration in mg L-1, were Ca (396 nm) 0.01, Ca 7 M. Thompson and M. H. Ramsey, Analyst, 1985, 110, 1413. (393 nm) 0.004, Cu (324 nm) 0.002, Mg (279 nm) 0.0001 and 8 A. S. Al-Ammar and R. M. Barnes, Spectrochim. Acta, Part B, Mg (285 nm) 0.007. For other analyte lines, the background 1998, 53, 1583. 9 A. S. Al-Ammar and R. M. Barnes, At. Spectrosc., 1998, 19, 18. equivalent concentration values were smaller. 10 X. Romero, E. Poussel and J. M. Mermet, Spectrochim. Acta, Part B, 1997, 52, 487. 11 J.-M.Mermet and J. C. Ivaldi, J. Anal. At. Spectrom., 1993, 8, 795. Conclusion 12 A. Boorn, M. S. Cresser and R. Browner, Spectrochim. Acta, Part Two spectral lines, an ionic and an atomic, of the same element B, 1980, 35, 823. can be used to correct for the diVerence in vapor loading and chemical composition between the sample and the calibration Paper 8/08389D J. Anal. At. Spectrom., 1999, 14, 801–807 807

ISSN:0267-9477    

DOI:10.1039/a808389d

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Determination of germanium at trace levels in environmental matrices by chloride generation-inductively coupled plasma atomic emission spectrometry† Silvia Farý�as and Patricia Smichowski* Comisio�n Nacional de Energý�a Ato�mica, Centro Ato�mico Constituyentes, Unidad de Actividad Quý�mica, Av. Libertador 8250, 1429-Buenos Aires, Argentina. E-mail: smichows@cnea.gov.ar Received 17th November 1998, Accepted 15th January 1999 A study was undertaken to evaluate the analytical capabilities of coupling chloride generation with inductively coupled plasma atomic emission spectrometry for the determination of germanium at trace levels.The method is based on the generation of germanium tetrachloride by concentrated hydrochloric acid in a continuous system and subsequent introduction of the gaseous analyte via the inlet tube of the plasma torch. After optimizing the conditions for GeCl4 generation as well as the plasma parameters, the method was applied to the determination of germanium in sediments, fly ashes and geological samples.A detection limit (3sblank) of 0.25 ng ml-1 was obtained. The precision (RSD) of the determination was 2.7% at a level of 100 ng ml-1 Ge and 4.6% for 10 ng ml-1 (n=10). The interferent eVect of various ions (As, Cd, Co, Cu, Fe, Ga, Hg, Ni, Pb, Sb, Se, Sn, Te, Zn) on the germanium signal was evaluated. Hydride generation and chloride generation of Ge were compared in terms of selectivity, sensitivity and precision.The vapour generation introduction technique has been suc- all cases. Brindle and Ceccarelli Ponzoni24 demonstrated that the use of ammonium peroxodisulfate eVectively overcame cessfully used in atomic spectrometric applications because it oVers several significant advantages in comparison with con- the interference from Al, As, Cu, Fe, Hg, Pb and Sn in the determination of Ge by HG-ICP-AES. Simultaneous signal ventional pneumatic nebulization methods.1–4 Reduction and volatilization of elements have the advantage of enhancing enhancement and reduction of interferences were achieved using L-cysteine, L-cystine, penicillamine, histidine and transport eYciency, approaching 100%, and separation of the analyte from the matrix that leads to improvement in the thioglycerol.30,31 The generation and introduction of volatile chlorides (As, selectivity in many instances.With vapour generation introduction techniques, detection limits can also be significantly Bi, Cd, Ge, Mo, Pb, Sn, Tl, Zn), fluorides (Ge, Mo, Re, U, V, W), b-diketonates (Co, Cr, Cu, Fe, Mn, Ni, Pd, Zn), metal improved, which is fundamental to obtaining the appropriate sensitivity for environmental analysis.In addition, gaseous carbonyls (Ni), volatile metal chelates (Fe, Co, Cr) and oxides (Os, Ru) into atomization cells is an interesting possibility to sample introduction promotes more eYcient plasma atomization, excitation and ionization of the analyte.The possibility reduce matrix interferences that has also been investigated.2,3 In this context, Skogerboe et al.32 developed a method based of preconcentration and speciation are other advantages to note. on the injection of diVerent chlorides, among them GeCl4, into a flame atomic absorption spectrometer and a microwave- Derivatization with sodium tetrahydroborate(III) has become the most popular method to determine elements such induced plasma. Guo and Guo33 proposed an interferencefree atomic fluorescence spectrometric method for the determi- as As, Bi, Ge, Hg, Pb, Sb, Se, Sn and Te in a diversity of matrices.5–10 The determination of Cu11 and Cd12–15 by hydride nation of Ge by utilizing the generation of GeCl4.The aim of the present work was to undertake a parametric generation has also been investigated. However hydride generation is not free from interferences. The most important optimization of the variables influencing the chloride generation (CG) of Ge prior to its determination in environmental interference occurs with transition metals,16–18 which reduce (or eventually enhance) the analytical signal.Also, concomit- matrices by ICP-AES and to investigate the potential of the chloride generation method to reduce or eventually eliminate ant eVects (gas phase atomization interferences) with other elements capable of generating volatile hydrides have been interferences. A comparison between CG and HG performances was also evaluated.reported in the literature.19–21 In the specific case of Ge, various hydride generationatomic spectrometric methods have been used by diVerent Experimental research groups.22–27 Inductively coupled plasma atomic emission spectrometry (ICP-AES) is a flexible detection technique Instrumentation and diVerent introduction systems such as hydride generation The instrumentation and optimized operating conditions for (HG) can be coupled to the system to improve sensitivity. the CG-ICP-AES system are summarized in Table 1.The Several acid media and reaction conditions22,25–29 have been continuous flow manifold used to generate the GeCl4 was tested to generate GeH4 (germane) but in spite of these various based on the use of a four-channel peristaltic pump (Gilson attempts, the problem of interferences was not overcome in M 312, Villiers Le Bel, France) and a laboratory-made generation cell similar to that described by Sturgeon et al.11 The sample and the carrier solutions were pumped at a flow rate †Presented at the Fifth Rio Symposium on Atomic Spectrometry, Cancu� n, Mexico, October 4–10, 1998.of 9 ml min-1 using acid resistant tubing. The GeCl4 generated J. Anal. At. Spectrom., 1999, 14, 809–814 809Table 1 Instrumentation and operating conditions NC, USA), equipped with 12 Teflon-PFA (perfluoroalkoxy) digestion vessels. ICP generation system HFP-2500 D (Plasma Therm, Kresson, NJ, USA) Torch Ames type, PT 1 (Plasma Therm) Reagents Nebulizer Meinhard type, TR-30-A3 (Precision Welding argon from AGA (Buenos Aires, Argentina) was Glassblowing, Englewood, CO, USA) Spray chamber Scott-type, double barrel, glass found to be suYciently pure for the Ge determination.All Monochromator VHR 1000, 1 m focal length (Jobin- reagents were of analytical reagent grade unless otherwise Yvon, Longjumeau, France). mentioned. Deionized water from Nanopure (Barnstead, Interferometric grating: 3600 Dubuque, IA, USA) was used throughout.A commercially lines mm-1 available 1000 mg l-1 Ge standard solution which contains Ge Integration time 4 s in the form of oxalate (Merck, Darmstadt, Germany) was Analytical wavelength Ge I: 303.906 nm used. Diluted working solutions were prepared daily by serial Plasma— dilutions of this stock solution. Nitric and hydrochloric acids Forward rf power 1.4 kW were prepared by distillation of analytical grade reagents. Frequency of rf generator 27.12 MHz Coolant (outer) gas flow 18 l min-1 A 3.0% (m/v) sodium tetrahydroborate(III) solution was rate prepared by dissolving NaBH4 powder (Baker, Phillipsburg, Auxiliary (intermediate) 0.6 l min-1 NJ, USA) in deionized water, stabilizing in 1.0% (m/v) NaOH gas flow rate solution (Merck) and filtering through Whatman No. 42 filter Sample (aerosol ) gas flow 0.8 l min-1 paper to eliminate turbidity. The solution was stored in a rate polyethylene flask at 4 °C.Diluted working solutions were Viewing height above 12 mm load coil prepared before use. All solutions containing the potential interferent ions studied Chloride generation— were prepared at the required concentrations by adding appro- Sample and reagents flow 9 ml min-1 rate priate amounts of stock solutions (Merck Titrisols) or their Sample acidity 5 M HCl+1.75 M HNO3 chloride salts in a mixture of 5 M HCl and 1.75 M HNO3. Carrier 12 M HCl Coil volume 750 ml Safety Tube size (samples and 1.5 mm (id) reagents) HCl is a corrosive reagent and should be handled with Hydride generation— appropriate safety precautions to avoid personal damage and Samples and reagents flow 9 ml min-1 corrosion of the equipment.rate Sample acidity 0.1 M HCl NaBH4 concentration 1.0% Sample preparation Coil volume 750 ml Tube size (samples and 1.5 mm (id) DiVerent aliquots homogenized samples were reagents) weighed into Teflon-PFA digestion vessels and decomposed using the acids and the microwave oven programme shown in Table 2.In order to standardize the acid content (optimized conditions) in the samples, the digests were slowly evaporated was swept out by Ar and introduced directly via the inlet tube of the plasma torch. The spray chamber was disconnected and to dryness and then dissolved in several millilitres of a mixture of 5 M HCl+1.75 M HNO3. The solutions were filtered and replaced with Tygon tubing to connect the phase separator with the torch.A schematic diagram of the instrumental transferred into 50 ml calibrated flasks and diluted to the mark with the same mixture. Three portions were weighed for assembly employed is shown in Fig. 1. Samples were digested using a CEM MDS-2000 microwave oven (CEM, Matthews, each sample. Fig. 1 Schematic diagram of the CG-ICP-AES coupling. 810 J. Anal. At. Spectrom., 1999, 14, 809–814Table 2 Microwave digestion programme for sediments, fly ashes and geological samples Reagents/ml Step 1 Step 2 Sample Sample mass/g HNO3 HF HCl Power/W t/min Power/W t/min Final volume/ml Sediments 0.5 5 4 1 630 30 0 0 50 Fly ashes 0.2 3 3 3 450 25 0 0 50 Geologicals 0.5 3 3 3 560 40 560 20 50 Procedure The acidified working solutions (5 M HCl+1.75 M HNO3) were merged at a Y-piece with the carrier (12 M HCl) and introduced into the chloride generator in a continuous flow system at 9 ml min-1 using a peristaltic pump.The GeCl4 generated was separated from the solution (Fig. 1) by means of a U-tube separator and swept by Ar (0.8 l min-1) into the bottom of the quartz torch (the spray chamber was disconnected). Tygon tubing was used for the connections. The system shown in Fig. 1 was also used for the hydride generation studies. The acidified working solution (0.1 M HCl) and the alkaline 1.0% NaBH4 solution (stabilized with 0.3% NaOH solution) were continuously introduced into the hydride generator and the evolved germane (GeH4) was fed into the plasma as described previously.No Ge signal was observed when blank solutions were Fig. 2 Influence of sample acidity on GeCl4 production: Ge, analysed. The analytical signals obtained were the average of 0.2 mg l-1; HCl (carrier), 12 M; flow rate (samples and carrier), five replicate measurements. Signals at the 303.91 nm Ge I line 5 mlmin-1. were processed using in-house software.34 Calibration was achieved using standards prepared in 5 M HCl+1.75 M HNO3 in the CG studies and 0.1 M HCl when the HG-ICP-AES concentrations, the analytical Ge signal decreases, probably coupling was tested.owing to the small fraction of Ge present as germanium The eVect of the interferents studied was calculated according tetrachloride. Hydrochloric acid concentrations higher than to eqn. (1), 8 M are not recommended because a continuous drop in the Ge signal is observed, which can be attributed to the loss of % variation=[(b-a)/b]×100 (1) analyte during the preparation of the samples because of the where a=Ge signal in the presence of interferent and b=Ge volatility of Ge.signal in the absence of interferent. Influence of carrier acidity on GeCl4 production. The eYciency Results and discussion of the generation of GeCl4 is also dependent on the carrier acidity. Fig. 3 depicts the eVect of HCl concentration on Ge It is crucial to optimize the chloride generation conditions for signal response with the acidity of the sample solution fixed the specific system in use.The design of the gas–liquid at 5 M HCl. A series of HCl solutions was prepared covering separator and the selection of the operating conditions deterthe range 8–12 M. As expected, Ge was found to give the best mine the performance of the coupling. Good long-term preresponse when concentrated hydrochloric acid was used as cision and accurate results also require strict control of the carrier because the complete formation of GeCl4 was achieved operating parameters. at this concentration.Chemical and physical parameters aVecting germanium chloride generation were optimized individually while other parameters were fixed at their optimum value in order to obtain maximum gas evolution. Unless otherwise noted solutions containing 0.2 mg l-1 Ge were used in the optimization process. Chemical parameters The chemical parameters investigated were: influence of sample acidity, carrier acidity and eVect of other acids on Ge signal.Influence of sample acidity on GeCl4 production. The rate at which germanium is volatilized is strongly dependent on the HCl concentration in the sample and in the carrier. In a first stage, the HCl concentration of the carrier was fixed (12 M) and the eVect of HCl on the acidified samples varying between 3 and 10 M was studied. The flow rate of acidified samples and carrier was 5 ml min-1. The eVect of acid concentration in the sample solutions is depicted in Fig. 2.An HCl concentration Fig. 3 Influence of carrier acidity on GeCl4 production: Ge, of 5 M was selected for subsequent work although a relatively 0.2 mg l-1; sample acidity, 5 M HCl; flow rate (samples and carrier), 5 mlmin-1. wide range of concentrations can be used. At lower HCl J. Anal. At. Spectrom., 1999, 14, 809–814 811EVect of other inorganic acids on germanium signal. A series of germanium-containing solutions in 5 M HCl was prepared and the eVect of other typical inorganic acids (phosphoric, nitric and sulfuric) on the Ge signal over a range of 2.5–20% was investigated.No eVect on the Ge signal was observed in the presence of H3PO4, while a strong depression was experienced when H2SO4 was tested. When HNO3 was evaluated, a significant increase in the Ge signal was observed. Fig. 4 illustrates the variation in Ge signal at various HNO3 concentrations. It is self-evident that 1.75 M is the most suitable HNO3 concentration because it enhances the analytical signal by approximately 100% and was thus used in further work.The ability to generate germanium tetrachloride in the presence of HNO3 is an important advantage to note because this acid is usually employed in sample digestion procedures. Physical parameters Fig. 5 Influence of sample and carrier flow rates on Ge signal: Ge, 0.2 mg l-1; sample acidity, 5 M HCl+1.75 M HNO3; HCl (carrier), 12 M. Physical parameters were optimized using a solution containing 0.2 mg l-1 of Ge with 12 M HCl as carrier and sample acidity The eVect of variation of the auxiliary gas flow rate between 5 M HCl+1.75 M HNO3. 0.2 and 2.3 l min-1 experienced a maximum for the Ge signal at 0.6 l min-1. Below this value a constant decrease was Wavelength selection. The ratio IGe5Ib (net germanium intenobserved that could be explained by an overheating of the sity5background intensity) was calculated for three Ge I inlet tube of the torch, producing premature decomposition emission lines (nm): 265.118, 275.459 and 303.906.The of the GeCl4. 303.906 nm wavelength was selected because it gave the best The response of the Ge signal with increasing coolant gas signal to background ratio. flow rate (12–20 l min-1) showed a maximum at 18 l min-1. If the coolant gas flow rate is used at lower values, spectral Optimization of sample and carrier flow rates. To evaluate interferences from molecular species, such as NO, can be the eVect of the pump speed, the acidified samples and the expected.carrier were pumped (1.5 mm id, PTFE tubing) over a range from 4 to 11 ml min-1. The best signals for Ge (Fig. 5) were Influence of forward power. The forward power was varied achieved with flow rates between 9 and 11 ml min-1. The between 1.2 and 1.6 kW and a maximum was experienced at pressures developed were too great at a flow rate higher than 1.4 kW. This power was used for further work. 11 ml min-1.However if better detection limits are not required, lower flow rates can be used in order to reduce EVect of glass balls on response. In order to obtain a better reagent consumption. vapour eYciency, minimize dead volume in the cell, increase the surface for vapour evolution and promote a smooth Optimization of Ar flow rates. The variation of one of the degassing of the waste solution,11 the U-tube of the generator three gas flow rates (coolant, auxiliary and sample) with the was filled with glass balls (2 mm diameter).The influence of other two fixed at a constant value was evaluated. the addition of glass balls is clearly evident in the calibration With the coolant and auxiliary gas flow rates fixed, the curves depicted in Fig. 6 where a significant increase in sensi- sample gas flow rate was varied over the range 0.2–1.0 l min-1 tivity is observed. The use of glass balls was adopted for and a maximum was observed at 0.8 l min-1. The variation further experiments.in sensitivity observed for the Ge signal can be attributed to diVerent excitation conditions experienced by the analyte when Interference study the Ar flow rate is changed. In addition, a balance between optimum transport eYciency and dilution is reached at this The selectivity of the method developed was evaluated by value. studying the eVect of several elements on the Ge signal with Fig. 4 Influence of HNO3 concentration on Ge signal: Ge, 0.2 mg l-1; Fig. 6 Influence of glass balls on Ge signal. Sample acidity, 5 M HCl+1.75 M HNO3; HCl (carrier), 12 M; flow rate (samples and sample acidity, 5 M HCl; HCl (carrier), 12 M; flow rate (samples and carrier), 5 ml min-1. carrier), 9 ml min-1. 812 J. Anal. At. Spectrom., 1999, 14, 809–814chloride and hydride generation, using the coupling shown in ments were also carried out under optimized conditions (Table 1) for the system depicted in Fig. 1. Fig. 1. All the tests were carried out under optimum operating conditions unless otherwise noted (Table 1).Variations over The analytical characteristics of the methods are depicted in Table 4. The detection limits calculated, following IUPAC ±5% in the analytical signal of Ge in the presence of foreign ions were taken as interferences. All test samples analysed rules,36 by ten measurements of the blank signal were 0.20 ng ml-1 (HG) and 0.25 ng ml-1 (CG). Compared with a contained 0.5 mg l-1 of Ge and the results are the average of three replicate measurements.conventional continuous nebulization (CN), hydride and chloride generation coupled to ICP-AES give a sensitivity Transition metals are serious interferents in the hydride generation technique as is described in the abundant literature increase of a factor of approximately three orders of magnitude (CN detection limit, 300 ng ml-1). The concentrations of existing on this subject. Welz and Melcher35 demonstrated that the hydride formed reacts with transition metals that have Ge over which the calibration curve was established (CG) in order to analyse the samples ranged from 5 to 200 ng ml-1. been reduced and precipitated in a colloidal form in the presence of tetrahydroborate. Under the experimental con- Parameters of the linear calibration graph are the following: a (intercept)=1.500, b (slope)=0.057 arbitrary units per ppb, ditions given in Table 1, the results of the interference study (Table 3) clearly show that elements that are serious inter- r (regression coeYcient)=0.998. Precisions were evaluated using two standards of Ge pre- ferences in the hydride generation of Ge do not interfere when the chloride generation is used. Only Ni severely interferes pared in 0.1 M HCl (HG) and 5 M HCl+1.75 M HNO3 (CG) containing 10 and 100 ng Ge ml-1.The values of RSD (%) with GeCl4 generation, resulting in a 30% suppression of the signal compared with that from a nickel-free solution.The obtained (n=10) are summarized in Table 4. Similar detection limits and precisions were obtained with both methods. The presence of a lower Ni concentration (50 mg l-1) also interferes with GeCl4 generation, resulting in a 10% reduction of the linear calibration ranges (HG and CG) span approximately three orders of magnitude. analytical signal of Ge. Guo and Guo33 have investigated the eVect of 20 mg ml-1 of Ni on the determination of Ge (100 ng ml-1) by GeCl4 generation-atomic fluorescence spec- Application to the analysis of environmental samples trometry (AFS).While they did not observe any interference, In order to demonstrate the reliability of the proposed method, a severe interference was observed in our work. It is not clear it was applied to the analysis of two reference materials: which eVects are responsible for these apparent diVerences. Antarctic Sediment (MURST ISS A1) and Stream Sediment When using hydride generation (optimized conditions), the (GBW 07307). Fly ashes and geological materials were also presence of 500 mg l-1 of Cu caused 95% suppression in the analysed.No reference standards for fly ashes and geological analytical signal of Ge, while when chloride generation was samples were available in our laboratory, so a recovery test evaluated, the interference was eliminated. The interference of was carried out by adding known amounts of Ge solutions Co and Fe in HG was also drastically reduced in CG.and applying the procedure previously described. The mean The lack of interference in the Ge signal observed when As recoveries of spiked Ge were in the range 96–103% depending was tested confirms that the volatilization of AsCl3 is slower on the sample analysed. than that of GeCl4. The results reported in this work for As Calibration was achieved using standards of Ge prepared are in good agreement with the observations of Guo and in 1.75 M HNO3+5 M HCl and from the data obtained no Guo.33 systematic errors due to the presence of any of the matrices The avoidance of the use of tetrahydroborate, as in the were observed.present CG method, has a direct influence on the reduction of The analytical results from environmental samples are pre- interferences and is a very important advantage to highlight. sented in Tables 5 and 6 (n=3). The interference study (Table 3) showed clearly that CG provides superior performance with respect to tolerance to transition metals and other hydride forming elements over Conclusions that of HG.The on line combination of chloride generation with ICP-AES provides an accurate and precise method for Ge determination Analytical performance Hydride and chloride generation were also compared in terms Table 4 Analytical performance of the couplings HG-ICP-AES and of their analytical performance. Hydride generation measure- CG-ICP-AES Parameter Hydride generation Chloride generation Table 3 EVect of potential interferents on the determination of Ge by HG-ICP-AES and CG-ICP-AES (Ge, 0.5 mg l-1) Detection limit (3s) 0.20 ng ml-1 0.25 ng ml-1 Precision (%)a 5.5 (for 10 ng ml-1) 4.6 (for 10 ng ml-1) Variation in Ge signal (%) 3.5 (for 100 ng ml-1) 2.7 (for 100 ng ml-1) Concentration/ an=10.Element mg l-1 HG-ICP-AES CG-ICP-AES AsIII 50 0 0 Table 5 Determination of Ge in environmental reference materials. CdII 50 -30 0 CoII 500 -95 6.3 Results expressed as mg g-1 Ge CuII 500 -95 0 FeIII 500 -95 -10 Sample Ge concentrationa GaIII 50 -22 3.9 HgII 50 -22 -5.6 Antarctic Sediment (MURST ISS A1)— Ge (found) 1.3±0.1 NiII 500 -95 -30 PbII 50 -43 2.0 Ge (informative concentration) 1±0.4 SbIII 50 -37 5.4 Stream Sediment (GBW 07307)— SeIV 50 -46 -6.9 Ge (found) 1.3±0.1 SnII 50 -7.0 5.3 Ge (certified ) 1.4±0.2 TeIV 50 -19 0 ZnII 500 -47 2.1 aMean value±standard deviation (n=3).J. Anal. At. Spectrom., 1999, 14, 809–814 813Table 6 Determination of Ge in fly ashes and geological samples. 7 S. Caroli, F. La Torre, F. Petrucci and N. Violante, Environ. Sci Pollut. Res., 1994, 1, 205. Results are expressed as mg g-1 Ge 8 P. Smichowski, Y. Madrid, M. B. de la Calle Guntin�as and C. Ca�mara, J. Anal. At. Spectrom., 1995, 13, 815. Samplea Concentrationb Recovery (%) 9 P. Fodor and R. Barnes, Spectrochim Acta, Part B, 1983, 38, 229. 10 M. O. Andreae, J.-F. Asmode�, P. Foster and L. Van’t dack, Anal. Fly ash 1 11.0±0.4 94 Fly ash 2 15.3±0.9 96 Chem., 1981, 53, 1766. 11 R. E. Sturgeon, J. Liu, V. J. Boyko and V. T. Luong, Anal. Chem., Granite 1.3±0.1 103 Greissen 6.7±0.3 102 1996, 68, 1883. 12 J. Narsito, J. Agterdenbos and D. Bax, Anal. Chim. Acta, 1991, aFly ashes and greissen samples were spiked with 40 ng ml-for 244, 129. recovery studies the granite sample was spiked with 10 ng ml-1 Ge. 13 H. Matusiewicz, M. Kopras and R. E. Sturgeon, Analyst, 1997, The tests were run in duplicate.bMean value±standard deviation 122, 331. (n=3). 14 A. Sanz-Medel, M. C. Valde�s-Hevia, Y. Temprano, N. Bordel Garcý�a and M. R. Ferna�ndez de la Campa, Anal. Chem., 1995, 67, 2216. at trace levels. The principal advantage of CG over HG is a 15 M. L. Garrido, R. Mun�oz-Olivas and C. Ca�mara, J. Anal. At. dramatic suppression of liquid phase interferences. The ability Spectrom., 1998, 13, 295. 16 B. Welz and M. Melcher, Anal. Chim. Acta, 1981, 131, 17. to generate GeCl4 in the presence of HNO3 is another advan- 17 J.Agget and G. Boyes, Analyst, 1989, 114, 1159. tage to highlight. 18 A. D’Ulivo, L. Lampugnani and R. Zamboni, Spectrochim. Acta, The comparison of chloride generation and continuous Part B, 1992, 47, 619. nebulization demonstrates that the optimized method pre- 19 J. Dedina, Anal. Chem., 1982, 54, 1982. sented in this study increases the detection power by three 20 K. Petrick and V. Krivan, Fresenius’ Z. Anal. Chem., 1987, 327, orders of magnitude. 338. 21 M. B. de la Calle-Guntin� as, R. Torralba, Y. Madrid, M. A. This method is simple and seems promising for coupling, Palacios, M. Bonilla and C. Ca�mara, Spectrochim. Acta, Part B, with only minor modifications, to other atomic spectrometric 1992, 47, 1165. detectors. 22 M. O. Andreae and Ph. N. Froelich, Jr., Anal. Chem., 1981, 53, 287. 23 J. R. Castillo, J. Lanaja and J. Azna�rez, Analyst, 1982, 107, 89. Acknowledgements 24 I. D. Brindle and C. M. Ceccarelli Ponzoni, Analyst, 1987, 112, 1547. The authors gratefully acknowledge Ruben E. Rodriguez, 25 T. Nakahara and T. Wasa, Microchem. J., 1994, 49, 202. Margarita Piccinna and Ariel Grillo for their helpful assistance 26 I. D. Brindle and X.-c. Le, Anal. Chem., 1989, 61, 1175. and Sergio Caroli for providing the certified reference mater- 27 I. D. Brindle, M. E. Brindle, X.-c. Le and H. Chen, J. Anal. At. ial MURST ISS A1. Juana Pedro is also gratefully acknowl- Specrom., 1991, 6, 129. edged for her valuable contribution in the hydride generation 28 K. Jin, Y. Shibata and M. Morita, Anal. Chem., 1991, 63, 986. measurements. This work is part of CNEA-CAC-UAQ pro- 29 P. Smichowski and J. Marrero, Anal. Chim. Acta, 1998, 376, 283. 30 I. D. Brindle, X.-c. Le and X.-f. Li, J. Anal. At. Spectrom., 1989, jects 97-Q-02–03 and 97-Q-02–05. 4, 227. 31 I. D. Brindle and X.-c. Le, Anal. Chim. Acta, 1990, 229, 239. References 32 R. K. Skogerboe, D. L. Dick, D. A. Pavlica and F. E. Lichte, Anal. Chem., 1975, 47, 568. 1 R. G. Godden and D. R. Thomerson, Analyst, 1989, 1257, 1137. 33 X.-w. Guo and X.-m. Guo, Anal. Chim. Acta, 1996, 330, 273. 2 X-P. Yan and Z-M. Ni, Anal. Chim. Acta, 1994, 291, 89. 34 R. E. Rodriguez, R. N. Garavaglia and D. A. Batistoni, ICP Inf. 3 H. Matusiewicz and R. E. Sturgeon, Spectrochim. Acta, Part B, Newsl., 1995, 20, 699. 1996, 51, 377. 35 B. Welz and M. Melcher, Analyst, 1984, 109, 573. 4 S. Clark and P. Craig, Mikrochim. Acta, 1992, 109, 141. 36 J. D. Winefordner and L. G. Long, Anal. Chem., 1983, 55, 713. 5 T. Nakahara, Appl. Spectrosc., 1979, 33, 206. 6 I. Aroza, M. Bonilla, Y. Madrid and C. Ca�mara, J. Anal. At. Paper 8/08981G Spectrom., 1989, 4, 163. 814 J. Anal. At. Spectrom., 1999,

ISSN:0267-9477    

DOI:10.1039/a808981g

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Determination of copper, iron and nickel in edible oils using emulsified solutions by ICP-AES† M. Murillo,a Z. Benzo,*b E. Marcano,b C. Gomez,b A. Garabotob and C. Marina aCentro de Quý�mica Analý�tica, Facultad de Ciencias, Universidad Central de Venezuela, Apdo. Postal 47102, Caracas 1041-A, Venezuela bCentro de Quý�mica, Instituto Venezolano de Investigaciones Cientý�ficas, IVIC, Apdo. Postal 21827, Caracas 1020-A, Venezuela Received 20th October 1998, Accepted 8th March 1999 The determination of copper, iron and nickel in edible oils was carried out using emulsion sample preparation followed by analysis by ICP-AES. Response surface methodology was applied in order to find the optimum emulsion and surfactant concentrations.The optimum amount of oil in the emulsion was in the range ca. 2–35% in most of the surfactants studied, except for Triton X-100, which showed a maximum response above 35% oil. The surfactant concentration in the emulsion varied between 0.5 and 9%.Good agreement was found between calibration curves for emulsified aqueous standards solutions and oil-in-water emulsions for most elements studied with Tween 80, ethoxynonylphenol and Triton X-100. The best agreement for all elements was shown when Tween 80 was used. Hence emulsified aqueous standard solutions could be used for the determination of these elements in emulsified edible oil samples using ICP-AES. Recoveries ranged from 90 to 110% for most of the elements studied, with relative standard deviations lower than 8%.the use of surfactants to enhance the sample transport process. Introduction Bertagnolli et al.8 reported the eVect of surfactants (Brij-35, The demand for the analysis of organic liquids most frequently Triton X-100 and UltraWet) on the analytical performance of arises in the determination of metal concentrations in organic ICP for the determination of Cu, Fe and Mn in samples from substances, e.g., crude oils, petroleum products, pharmaceut- a steel processing operation.They concluded that surfactant ical or industrial organic samples and food products. A great eVects on sample transport are modest. Varying surfactant deal of research has been devoted to attempts to discover why concentrations produced no substantial changes in performsome fats and oils undergo changes more rapidly than others, ance and the presence of surfactants had no eVect on the what the changes are, what causes the changes and how they sensitivity of the instrument.On the other hand, Murillo et al.9 can be controlled. The quality assessment of fats with regard reported emulsion sample introduction for the determination to freshness, keeping properties and storability can be carried of metals in lubricating oils by ICP-AES and Goncalves et al.10 out by the determination of metals. Trace of heavy metals in carried out the determination of metals in used lubricating edible oils are known to have an eVect on the rate of oil oils by AAS using emulsified samples.In these studies, good oxidation. Particularly copper and iron seriously aVect oil agreement was found between calibration curves of aqueous quality and lead, arsenic and mercury are the subject of food and emulsified standard solutions. legislation. Information about which surfactant works best and the Analyses of edible oils for diVerent elements have been optimum concentration of surfactant to be used is sparse.The carried out by electrothermal atomic absorption spec- solubility of non-ionic surfactants in water can usually be used troscopy1–3 and dc plasma emission spectroscopy.4 Many as a guide in approximating their hydrophile–lipophile balance diVerent sampling procedures have been used for the analysis (HLB) and their usefulness. By choosing surfactants with of oils by inductively coupled plasma atomic emission spec- HLB values appropriate for emulsifying any given oil system, troscopy (ICP-AES).These include dilution with a suitable trial-and-error eVorts are reduced and optimum performance organic solvent, dry and acid digestions and extraction is usually obtained rapidly. Hydrophilic surfactants are water methods, which are prone to losses and contamination and soluble and are used for solubilization. Surfactants in the HLB are time consuming. The direct analysis of organic solutions range 9.6–16.7 are considered hydrophilic, and tend to form by ICP methods encounters several diYculties.In general, the oil-in-water (o/w) emulsions. analytical method has a lower performance in organic media We studied the use of o/w emulsions in the determination than aqueous solutions. On the other hand, calibration is more of Cu, Fe and Ni in edible oils by ICP-AES. To our knowledge, diYcult in organic media. An optional method of sample trace metal analyses of edible oils using emulsified samples by introduction into the plasma is to use systems such as emul- ICP-AES have not been reported.Copper, nickel and iron sions. One of the principal advantages of the use of emulsions were chosen since these metals have a negative influence on in the ICP-AES determination of metallic elements in organic the oxidative stability of edible oils and fats. Hydrogenation matrices is that aqueous solutions can be used for the prep- of edible oils and fats has been carried out using nickel aration of the standards.Many researchers5–7 have suggested catalysts. Copper and iron are present from processing equipment. DiVerent non-ionic surfactants were used: ethoxynonylphenol, Tween 21 and 80 and Triton X-100. They were chosen †Presented at the Fifth Rio Symposium on Atomic Spectrometry, Cancun, Mexico, October 4–10, 1998. on the basis of availability and according to the moles of J. Anal. At. Spectrom., 1999, 14, 815–820 815ethylene oxide present in the molecule, except for ethoxynonyl- Optimization of ICP Conditions phenol and Triton X-100.The justification for the use of these In order to evaluate the influence of the operating parameters two is that even though they have similar HLB values and the on the intensity ratio of the Mg II 280.270 nm line to the Mg same number of moles of ethylene oxide, their chemical I 285.213 nm line, a two-level and a five-factor factorial study structures are diVerent. was designed. The five factors considered were power, nebulizer argon gas flow rate, intermediate argon gas flow, solution flow Experimental rate and viewing height.The optimum conditions found are given in Table 1. Instrumentation Design A Perkin-Elmer (Norwalk, CT, USA) Model ICP Optima 3000 inductively coupled plasma emission spectrometer A five concentration level design11 was chosen. Two factors equipped with Perkin-Elmer Model AS 90 autosampler was are examined: oil concentration (from 0 to 50% w/w) and used.The instrumental and operational parameters are surfactant concentration (from 0 to 10% w/w). The response described in Table 1. A standard ‘demountable’ type quartz chosen was the emission intensity of each analyte under study. plasma torch was used throughout. The i.d. of the alumina The concentration range chosen for both factors was based injector was 1.5 mm. A 10-roller peristaltic pump was used to on a literature review. This design is orthogonal. feed the nebulizer with the sample solution.All functions of the plasma were computer controlled. Results and discussion Reagents In order to find the optimum surfactant and o/w emulsion concentrations to be used in the subsequent parts of this work, All reagents were of the highest available purity. Ultrapure response surface methodology was applied. This methodology water was obtained from aMilli-Q system (Millipore, Bedford, has been applied to find values of the inputs which yield a MA, USA).Multi-element working standard solutions were maximum for a specific response and find how close this freshly prepared as required from individual element stock response surface is to this maximum. Fig. 1 and 2 show a standard solutions (BDH, Poole, Dorset, UK and Aldrich, three-dimensional representation of the response as a function Milwaukee, WI, USA). Working standard solutions used for of the ors, for nickel. An additional representation of calibration purposes were prepared by suitable dilution of the the same response for each case is also shown to see more stock standard solutions.All glassware was cleaned in nitric clearly the magnitude of the maximum response. Similar acid prior to use. High purity (99.95%) argon was utilized. response surfaces were obtained for copper and iron. The surfactants used were ethoxynonylphenol from Etoxyl (Maracaibo, Venezuela), Tween 21 and 80 from Aldrich and Ethoxynonylphenol Triton X-100 from J.T. Baker (Philipsburg, NJ, USA). Copper determined in the emulsion formed with this surfactant Sample presents maximum emission intensities when the composition of the o/w emulsion falls between 0 and 35% w/w with 7.5–10% A commercial sunflower oil sample was used in this work. w/w ethoxynonylphenol. Iron shows a maximum emission intensity for ca. 2.5–30% w/w o/w emulsion with ca. 8.5–10% Emulsion preparation w/w ethoxynonylphenol. Maximum emission intensity for Oil-in-water emulsions were prepared by weighing the sun- nickel is obtained at ca. 2.5–30% o/w emulsion with 8–10% flower oil samples together with the respective surfactant and w/w, ethoxynonylphenol. according to the scheme dictated by the design. They were left to stand overnight and then mechanically shaken for 30 min Triton X-100 prior to analysis. A very narrow zone of maximum emission intensities was Oil samples emulsified with ethoxynonylphenol showed obtained from the two variables studied with the emulsion fairly good stability after manual shaking.Emulsions formed formed with this surfactant. Copper, iron and nickel maximum with Triton X-100 were unstable. Tween 80 and Tween 21 mix emission intensities showed the same optimum zone, with o/w well with the vegetable oil under study and the emulsion emulsion concentration ca. 45–50% w/w and up to ca. 0.5% formed looks more stable after manual shaking. w/w Triton X-100. However, it can be said that reasonably good emission intensities were obtained for diVerent sets of values of the variables studied (see white region in Fig. 2). Table 1 ICP operating parameters Rf generator 40 MHz Tween 80 Operating power 1400 W Emulsions formed with Tween 80 showed maximum emission Nebulizer Meinhard Spray chamber Scott Type intensities under the following optimum conditions: ca. Sample delivery Peristaltic pump 2.5–45% w/w o/w emulsion and ca. 6.5–10% w/w surfactant Pump uptake rate 1.0 ml min-1 for copper, ca. 2.5–40% w/w o/w emulsion and 7.5–10% w/w Nebulizer flow rate 0.50 l min-1 surfactant for iron and ca. 2.5–35% w/w o/w emulsion and Plasma gas flow rate 15 l min-1 7.5–10% w/w surfactant for nickel. Auxiliary gas flow rate 1.7 l min-1 Response surface methodology could not be applied for the Observation height above r.f. coil 5 mm Background correction Automatic surfactant Tween 21 owing to the lack of suYcient material. Integration time 200 ms For calibration studies, we adopted the same optimum con- Working wavelength— ditions as found for Tween 80.Fe II 259.940 nm Cu I 327.396 nm Calibration Cu II 224.700 nm Ni I 232.003 nm The analytical interest of this study was to check the possibility Ni II 231.604 nm of using aqueous calibration solutions to quantify edible oil 816 J. Anal. At. Spectrom., 1999, 14, 815–820Fig. 1 Three and two-dimensional representations of the emission intensity for the Ni I 232.003 nm line as a function of emulsion and surfactant concentrations.Fig. 2 Three and two-dimensional representations of the emission intensity for the Ni I 232.003 nm line as a function of emulsion and surfactant concentrations (continuation of Fig. 1). J. Anal. At. Spectrom., 1999, 14, 815–820 817Fig. 4 Comparison of the calibration curves for Cu, Ni and Fe in the three systems using Triton X-100 (0.5%) as emulsifier. The application of the test (at the 95% confidence level ) shows that the calibration curves of the aqueous standards Fig. 3 Comparison of the calibration curves for Cu, Ni and Fe in the solutions containing ethoxynonylphenol (9%) and those resulting three systems using ethoxynonylphenol (9%) as emulsifier. from the o/w emulsions did not diVer significantly when copper and iron were determined at the 224, 327 and 259 for Cu II, Cu I and Fe II lines, respectively (Fig. 3). For nickel, the samples, once their accuracy for such purposes has been established experimentally.Thus, from the response surface results were significantly diVerent. The results using Triton X-100 show that the o/w emulsion did not diVer from the methodology results (estimation of the best o/w and surfactant concentrations), several sets of calibration curves for the aqueous standard solutions containing Triton X-100 (0.5%) in analyses using Cu I, Fe II and Ni II (Fig. 4). Systematic surfactant in an aqueous matrix and the surfactant in an oil matrix were compared with those given by standard aqueous diVerences among the results were obtained for the calibration curves obtained with the surfactant Tween 21 (Fig. 5). solutions. The statistical test in which the regression lines are used for comparing analytical methods12 was applied in order The results for the surfactant Tween 80 (Fig. 6) show that the calculated slopes and intercept for the aqueous standards to find evidence (if any) for systematic diVerences between the sets of results.For this, the calculated slope and intercept of solutions containing Tween 80 (9%) and those of the o/w emulsions did not diVer significantly when copper, iron and the calibration curves should not diVer significantly from the ‘ideal’ values of 1 and 0, respectively, when two diVerent nickel were determined using the 224, 327, 259, 232 and 231 nm Cu II, Cu I, Fe II, Ni I and Ni II, lines, respectively. conditions are examined. In this case, we compared the calibration curves obtained from the emulsified aqueous stan- From the above results, it can be concluded that among the calibrations studied, the aqueous calibration curve obtained dards and the o/w emulsions (Table 2).Fig. 3–6 show calibration curves obtained from aqueous standards, aqueous using the surfactant Tween 80 shows the best agreement with the emulsified standard solutions for all the elements con- standards with the surfactant added at the optimum estimated concentration and the emulsified oil samples.sidered. The statistical study of the solutions containing the 818 J. Anal. At. Spectrom., 1999, 14, 815–820Fig. 6 Comparison of the calibration curves for Cu, Ni and Fe in the three systems using Tween 80 (9%) as emulsifier. Surfactants with HLB in the range 9.6–16.7 are considered Fig. 5 Comparison of the calibration curves for Cu, Ni and Fe in the hydrophilic (water-soluble), and tend to form o/w emulsions.13 three systems using Tween 21 (9%) as emulsifier.The solubility of non-ionic surfactants in water can usually be used as a guide in approximating their hydrophile–lipophile balance and their usefulness. It has been reported13 that a surfactant ethoxynonylphenol showed that they can also be used to quantify the o/w emulsions for copper with the 224 nm good starting point for the emulsifier concentration is 3% for a system with low solids to as much as 10% for a system with emission line and for iron. However, nickel did not exhibit this behavior.high solids. Based on these, Triton X-100, ethoxynonylphenol, Tween 80 and Tween 21 with HLB values of 13.5, 13.9, 15.04 The presence of Tween 21 did not allow the use of aqueous standards solutions for calibration purposes. It can also be and 13.3, respectively, were chosen, depending also on availability. On the other hand, we also intended to establish noted that there was no diVerence in the response when either the atomic or ionic lines were used.whether there is any correlation between the results obtained and the number of ethylene oxide units present in the molecule. The results obtained in this work deserve some comments about the choice of the surfactants used. The HLB number of It has been reported14 that the increment of the number of ethylene oxide moieties in the molecule of any type of surfac- each surfactant gives an indication of good emulsification. By choosing surfactants with HLB values appropriate for emul- tant increases their solubility in water and decreases their sensitivity to pH changes.Triton X-100, ethoxynonylphenol, sifying any given oil system, trial-and-error eVorts are reduced and optimum performance is usually obtained rapidly. Tween 80 and Tween 21 have 9.5, 9.5, 20 and 4 mol of ethylene J. Anal. At. Spectrom., 1999, 14, 815–820 819Table 2 Statistical tests: o/w emulsion versus aqueous emulsion With the exception of nickel, at the lowest spiked concentration (and at the two emission lines), the recoveries for all spikes, Surfactant Line/nm Slope±t/sa Intercept±t/sa 400, 800 and 2000 ng g-1, are in the 90–110% range for all the elements considered with RSDs lower than 8%.Tween 80 Cu I 327 1.09±0.05 -34.32±306.96 Cu II 224 0.94±0.08 -79.82±165.47 Ni II 231 1.05±0.04 -54.98±97.75 Conclusion Ni I 232 0.95±0.06 -0.82±43.25 Fe II 259 1.05±0.04 151.08±297.45 Metal analysis of emulsified edible oils using aqueous standard Tween 21 Cu I 327 0.83±0.03 215.35±195.14 solutions for calibration purposes proved to be a promising Cu II 224 0.81±0.03 102.08±71.44 approach.In view of the results obtained, it can be said that Ni II 231 0.86±0.03 15.23±80.56 the surfactant Tween 80, when present in the aqueous solu- Ni I 232 0.83±0.03 7.94±23.63 Fe II 259 0.83±0.05 330.47±313.72 tions, showed a good statistical correlation with the emulsified Triton X-100 Cu I 327 0.99±0.07 226.93±296.60 oil solutions.This surfactant produced the best emulsified Cu II 224 1.20±0.04 -10.39±66.06 solutions. Hence it can be used for the determination of the Ni II 231 1.04±0.06 97.90±149.75 metals considered in this work in emulsified edible oil samples. Ni I 232 1.16±0.05 9.19±33.96 Fairly good results were obtained for ethoxynonylphenol and Fe II 259 1.03±0.04 316.40±216.56 Triton X-100, but their emulsions were not so stable. Ethoxynonylphenol Cu I 327 0.91±0.03 103.26±176.74 Cu II 224 0.94±0.07 89.88±161.41 Furthermore, the recoveries obtained, 90–110% for most of Ni II 231 0.93±0.04 -40.77±123.21 the elements, except nickel at both the emission lines (68 and Ni I 232 0.91±0.05 -1.00±37.72 60%) at the lowest spiked concentration, show that the pro- Fe II 259 0.95±0.09 -237.43±677.74 posed method is accurate for this type of determination.aTheoretical slope=1; theoretical intercept=0; theoretical t=2.57 (95% confidence level ). Acknowledgement The authors gratefully acknowledge the valuable support of Table 3 Recoveries (%) and relative standard deviations (%) (in parthe program BID-CONICIT through the project QF-10 and entheses) for spiked sunflower oil Dr.G. Urlina, Centro di Fisica, IVIC, for his initial support. Spiked concentration References Element 400 ng g-1 800 ng g-1 2000 ng g-1 1 F. J. Slikkerveer, A. A. Braad and P. W. Hendrikse, At. Spectrosc., Fe (259) 90.0 (6) 94.4 (1) 108.3 (1) 1980, 1, 30. Cu (224) 110.4 (3) 105.2 (3) 106.6 (7) 2 R.Calapaj, S. Chiricosta, G. Saija and E. Bruno, At. Spectrosc., Cu (327) 96.0 (4) 94.3 (3) 101.2 (7) 1988, 9, 107. Ni (231) 68 (24) 103.8 (7) 104.1 (2) 3 A. M. Nash, T. L. Mounts and W. F. Kwolek, J. Am. Oil Chem. Ni (232) 59.7 (28) 95.3 (6) 97.5 (3) Soc., 1983, 60, 811. 4 A. J. Dijkstra and D. Meert, J. Am. Oil Chem. Soc., 1982, 59, 199. 5 A. P.M. De Win, J. Anal. At. Spectrom., 1988, 3, 487. oxide, respectively. The HLB system and the number of ethy- 6 J. C. Ivaldi and W. Slavin, J. Anal. At. Spectrom., 1990, 5, 359. lene oxide moieties present in the molecule seem to be a 7 C. G. Millward and P. D. Kluckner, J. Anal. At. Spectrom., 1991, characteristic that may be indicative of the results obtained. 6, 38. 8 J. A. Bertagnolli, D. L. Neylan and D. D. Hammargren, At. Tween 80 has the highest HLB value among the surfactants Spectrosc., 1993, 14, 4. studied. It also has the highest number of ethylene oxide units. 9 M. Murillo, A. Gonzalez, A. Ramirez and N. Guillen, At. We are unable to explore the emulsion theory to explain in Spectrosc., 1994, 15, 90. more detail the results obtained. 10 I. M. Goncalves, M. Murillo and A. M. Gonzalez, Talanta, 1998, 47, 1033. Accuracy of the proposed method 11 R. G. Brereton, Analyst, 1997, 122, 1521. 12 J. C. Miller and J. N. Miller, Statistics for Analytical Chemistry, In order to check the accuracy of the proposed methodology, Wiley, New York, 1984, p. 102. spike and recovery experiments were carried out at 400, 800 13 ICI Americas, ICI Surfactants Web Site: and 2000 ng g-1 for each element. The results are given in http://www.surfactant.com/HLBSystem.html. 14 M. De la Guardia, Applicaciones Analý�ticas de los Agentes Table 3. The values reported are the average recoveries for Tensoactivos, Universidad de Valencia, Valencia, Spain, 1980. five samples with relative standard deviations (RSDs). Paper 8/08159J 820 J. Anal. At. Spectrom., 1999, 14, 815–8

ISSN:0267-9477    

DOI:10.1039/a808159j

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Determination of chromium in urine by electrothermal atomic absorption spectrometry using diVerent chemical modifiers† J. L. Burguera,*a M. Burguera,a C. Rondon,a L. Rodrý�guez,a P. Carrero,a Y. Petit de Pen�aa and E. Burguerab aIVAIQUIM (Venezuelan Andean Institute for Chemical Research), Faculty of Sciences, University of Los Andes, Apartado Portal 542, Me�rida 5101-A, Venezuela bFaculty of Odontology, Department of Preventive and Social Odontology, University of Los Andes, Apartado Postal 542, Me�rida 5101-A, Venezuela Received 23rd October 1998, Accepted 8th February 1999 The eVects of the chemical modifiers Eu, Mg(NO3)2, Pd, Eu–Pd and Ni and the use of longitudinally (with deuterium lamp background correction) and transversally (with Zeeman eVect background correction) heated atomizers on the determination of chromium in urine samples were studied.When longitudinally heated atomizers were used, the background absorption was too high for the addition of most species, such as Mg(NO3)2, Pd, Eu–Pd and Ni, rendering its use unsuitable for the determination of chromium in urine. The addition of Eu led to lower background signals.However, when longitudinally heated atomizers were used larger background absorption signals were observed on addition of a modifier. In contrast, with transversally heated atomizers the background absorption signals were low and were not influenced by the addition of a modifier, except when Mg(NO3)2 was added.In this case, the atomization signal shapes and the sensitivity were improved. The characteristic masses (m0) and limits of detection (3s) were 2.2 and 3.3 pg and 0.02 and 0.04 mg Cr l-1 for the longitudinally (using Eu as modifier) and transversally [using Mg(NO3)2] heated atomizers, respectively. Recovery studies and analysis of standard reference materials certified for chromium were performed to asses the accuracy. The results for the determination of chromium in real samples with both background correction procedures agreed well with a precision between 0.8 and 2.5%.Both procedures can be recommended and the choice will depend on instrument availability. Chromium is an essential element for all vertebrates. In during sample collection or diYculties in the determination of humans, it plays a role in the metabolism of glucose and some the element at low concentrations in a complex matrix. lipids (mainly cholesterol ). Chromium deficiency is associated Electrothermal atomic absorption spectrometry (ETAAS) is with cardiovascular diseases and diabetes, while excessive probably the most popular technique today for the determiamounts of the element, particularly in the more toxic Cr(VI) nation of the low concentrations of chromium expected in valence state, are detrimental to health as it may be involved clinical samples.in the pathogenesis of some diseases such as lung and gastroin- Several workers have used diVerent atomization techniques testinal cancer.(wall, platform3 and probe4)5,6 using diVerent atomization The main sources of chromium in the environment include surfaces (molybdenum,4,6 electrographite,5 pyrolytic graphitewaste chromates from electroplating baths, corrosion inhibi- coated graphite3–5 and totally pyrolytic graphite5), the use tors from heat exchange systems and trade eZuents from of diVerent chemical modifiers [Ca(NO3)2+Mg(NO3)2,7 tanning leather processes.Chromium in ambient air originates Mg(NO3)2,8 thiourea4] and diVerent background correction from industrial sources, particularly ferrochrome production, methods (tungsten–halogen lamp,9–11 deuterium lamp (DBC) ore refining, chemical, coal-fired power plants, cement- and Zeeman eVect (ZBC)5,9,12,13 correction) in order either to producing plants, refractor processing and combustion of improve the precision,9 to lower the detection limit5 or to fossil fuels.1 However, the dissolution of chromium from decrease matrix interferences in the determination of chromium stainless steel, which is widely used in the food industry, is in urine.Some workers have proposed direct procedures using probably the main source of chromium contamination of food. fast heating and pyrolytic tubes with14 and without back- On the other hand, the routes by which chromium enters the ground (BG) correction.15 The success of the determination body are the digestive tract, respiratory system and skin.It is of chromium by ETAAS16 seems to rely on (i) an accurate excreted principally in the urine and in small amounts in the setting of the pyrolysis temperature, (ii) an eVective backhair. Hence, it is clear that monitoring of waters and body ground correction method, especially ZBC, and (iii) the use fluids (especially urine and blood serum) for chromium is of an appropriate chemical modifier, at least the addition of essential for the control of nutritional deficiencies and perhaps species that would promote the formation of Cr2O317 either prevent its toxic eVects in cases of occupational exposure.in the sample pre-treatment or directly introduced into the The levels of chromium reported in human urine are very atomizer. low in healthy subjects, ranging from 0.2 to more than We have found that some isomorphous metals are more 12 mg l-1 with a urinary excretion below 10 mg d-1.2 This wide eYcient chemical modifiers for the determination of lead in range could be attributed to either sample contamination whole blood and urine (Eu),18 nickel in saliva (Pd–Lu),19 strontium in bones and whole blood (La)20 and iron in fingernails (Lu).21 Therefore, this work presents a comparative †Presented at the Fifth Rio Symposium on Atomic Spectrometry, Cancu� n, Mexico, October 4–10, 1998.study of several potential chemical modifiers, such as Eu, J. Anal. At. Spectrom., 1999, 14, 821–825 821Mg(NO3)2, Pd and Ni(NO3)2, alone or combined, for the grated absorbance obtained from 0 to 10 (with the ZBC) and from 0 to 20 (with the deuterium lamp) mg l-1 Cr solutions.determination of chromium in urine using longitudinally (LHGA) and transversally (THGA) heated graphite atomizers The average absorbance values of three or four injections were obtained in each case. Pyrolysis temperatures from 800 to with DBC and ZBC systems. 1600 °C and atomization temperatures from 1800 to 2600 °C were tested with the diVerent instrumentation.Experimental Instrumentation Results and discussion The experiments were carried out using two diVerent Perkin- Comparison of diVerent atomizers Elmer (Norwalk, CT, USA) atomic absorption spectrometers: Models 2100 (PE 2100) and 4100 ZL (PE 4100), with LHGA Similar symmetrical atomization signal shapes with low BG and THGA and with DBC and ZBC, respectively. were observed for chromium aqueous solutions using platform Instrumental control and data processing were accomplished atomization (PA) and both types of atomizers in the absence with (i) an IBM computer, Model 50Z, running Perkin-Elmer’s of any modifier.PA using LHGA with DBC provided much property software (version 9.1) and (ii) an Epson personal higher sensitivity than using THGA with ZBC. However, computer, Model EL 486UC, through Perkin-Elmer 4100PC when LHGA with DBC was used for the determination of software (version 7.3), under Gem Desktop (version Gem/3) chromium in urine samples, larger and lower analytical signals for the PE 2100 and PE 4100 ZL instruments, respectively.In were obtained with non-specific absorbances and significant both cases, a Perkin-Elmer chromium hollow-cathode lamp, over-correction eVects (Fig. 1A,a). When no BG correction is pyrolytic graphite-coated graphite tubes and pyrolytic graph- used (Fig. 1A,b), the signal is very diVerent to the dotted BG ite-coated graphite platforms were used.Integrated absorbance curve. This indicates that the BG at the chromium line, as (peak area) values, peak profiles and statistical data were measured with the hollow-cathode lamp, is diVerent to the printed with an Epson Model LX-810 printer. BG over the spectral bandpass of the monochromator, which indicates that the DBC would not properly correct. The Reagents and standard reference materials uncorrected abs separated into three peaks. The first two peaks separated by a valley are probably due to back- All chemicals were of analytical-reagent grade, unless stated otherwise.The glassware and polypropylene containers were ground absorbance originating from organic moieties present in the sample and the last peak is due primarily to the kept in 10% v/v HNO3 for least overnight and then rinsed with 1% v/v HNO3 three times and subsequently six to eight inorganic content of the matrix. However, when PA with ZBG was used, similar signal tracings were obtained with no excess- times with water before use.A 1.000 g l-1 certified chromium solution, as CrCl3, from Merck (Darmstadt, Germany) was ive BG signals (Fig. 1B) with and without BG attenuation, presumably because the THGA allows atomization in an used. Working standard solutions were prepared daily by appropriate dilution of the stock standard solution with isothermal environment. The diVerences observed previously could be primarily due Milli-Q (Millipore, Bedford, MA, USA) purified water acidi- fied to a concentration of 1% v/v HNO3.Nitric acid was of to the diVerent designs of the two atomizers used.5,22–24 The PE 2100 atomizers are 28 mm long whereas the PE 4100 Suprapur grade from Merck. Solutions of 1000 mg l-1 Pd, Eu, Mg and Ni were prepared by dissolving Pd(NO3)2 (pro atomizers are only 17 mm long. Also, the mode of heating of the LHGA is diVerent (end-side-heated), which provides tem- analysi, Merck), Eu2O3 (99.9%, BDH, Poole, Dorset, UK), Mg(NO3)2·6H2O (Baker, Deventer, The Netherlands) and perature gradients within the tubes during heating, whereas the thermal conditions in the THGA tube remain unchanged.Ni(NO3)2 (BDH) as described previously.19 The standard reference materials (SRMs) Seronorm Trace Thus, although THGA significantly reduces or eliminates condensation of the sample matrix components and ‘memory’ Elements in Urine from Nycomed (Oslo, Norway) and SRM 2670 Freeze-dried Urine from NIST (Gaitherburg, MD, USA) eVects and improves the atomization eYciency for refractory elements, such as chromium, longer tubes with larger sizes do were used to test the accuracy of the results.not permit eYcient elimination of matrix components in the Sampling atomization step. Moreover, the measurements are complicated by diYculties due to (a) DBC at the relatively long Urine samples (24 h) were collected from healthy laboratory chromium resonance wavelength (357.9 nm), where the emis- volunteers and stainless-steel factory workers in sterile acidsion intensity of deuterium lamps is low and the emission of washed disposable polypropylene containers.Sub-samples of elements of the matrix, e.g., urine, can be high25 and (b) the 10 ml were pipetted from these samples into 10 ml capped, scattering of the tube radiation by particulate matter from the acid-washed, high pressure polyethylene test tubes and stored matrix (toward the detector), which causes BG correction at 4 °C.The samples were analyzed directly without dilution. problems. This discourages the use of the PE 2100 tubes for Precautions were taken to avoid extraneous chromium conthe analysis of samples containing high levels of saline or tamination of the samples. All plastic sample containers, organic matrix, unless an eYcient chemical modifier is used. glassware and autosampler cups were soaked in 10% v/v nitric In order to obtain more data that could provide more acid overnight and then rinsed thoroughly with ultrapure convincing arguments for the choice of the atomizer, the BG water (Milli-Q) until the washings were ‘free of chromium’ as and specific signals for chromium in a urine sample, with and tested by the proposed ETAAS technique using either kind of without BG correction, were evaluated (Table 2).The results atomizer. indicated that the integrated absorbance signals were not aVected by the use of ZBC, producing readings statistically Procedure indistinguishable (p>0.01) from one other.This further con- firms the above observations that Cr was atomized in the PE For the choice of the pyrolysis temperature (Tpyr), the atomization temperature and optimum mass of the modifier, 20 ml of 4100 graphite tubes in an environment free from interferences. However, diVerent sensitivities with high BG signals were the chromium solution (40 pg) or sample and 10 ml of the modifier solution were injected sequentially into the L’vov obtained for Cr signals when the PE 2100 atomizer was used.This may indicate greater spreading of the chromium species platform. The temperature program in Table 1 was followed. Calibration graphs were constructed by evaluating the inte- in the relative larger size of the graphite tube cavity. It is 822 J. Anal. At. Spectrom., 1999, 14, 821–825Table 1 Temperature program for the graphite furnace using Model 2100 and 4100 instruments Argon flow rate/ Temperature/°C Ramp time/s Hold time/s mlmin-1 Step 2100 4100 2100 4100 2100 4100 2100 4100 Drying 90 110 5 1 10 20 300 250 Drying 120 130 10 5 10 30 300 250 Pyrolysis 1500a 1400a 5 10 20 20 300 250 Atomization 2500 2300 0 0 5 5 0 0 Cleaning 2600 2400 1 1 3 2 300 250 Cooling 20 20 3 1 5 5 300 250 aTpyr.Other variables common for both instruments: l=357.9 nm, integration time=5 s, measurement mode=peak area, hollow-cathode lamp current=18 mA and slit width=0.7 nm.and temperatures predetermined by the Perkin-Elmer software. When LHGA with DBG correction was tested, the integrated absorbance signals in the absence and presence of most modifiers showed imperceptible BG signals for the determination of chromium in aqueous solutions. However, these BG signals were too high (Fig. 2) when chromium was determined in urine samples, except when Eu was added. This diYculty in the measurements could be due to the relatively long chromium wavelength (357.9 nm), where the emission intensity of deuterium lamps is low and the emission of elements of the matrix is high.19 DiVerent amounts of Mg(NO3)2 (up to 10 mg of Mg), Pd (up to 1 mg), Ni (up to 2 mg) and Eu–Pd (up to 2 mg of each) did not stabilize Cr.In fact, the addition of these elements produced erratic and noisy peaks and the BG was always high. Europium was the only element that stabilized Cr. The optimum amount of Eu was 0.4 ng, because larger Fig. 1 Absorbance profiles for Cr in a urine sample.(A) Using PA amounts produced peaks with tailing and decreased sensitivity. atomization, (a) with and (b) without DBC; and (B) using PA When chromium was determined in urine samples using atomization, (a) with and (b) without ZBC. Other experimental THGA with ZBC, the BG signals were considerably reduced. conditions as specified in Table 1. The integrated absorbance signals were not aVected by the addition of most modifiers, except when 7–10 mg of Mg were Table 2 Comparison of integrated absorbances and background readingsa obtained for chromium in a urine sample using the PE 2100 added.In this case, the signals increased for an aqueous (with LHGA and DBC) and PE 4100 (with THGA and ZBC) chromium solution and for a urine sample (Fig. 3A,b). Also, atomizers, with and without background correction the relatively low BG signal which appeared when chromium was measured in urine samples in the absence of any modifier Atomization technique A s BG was greatly minimized (Fig. 3B and C). In this case, the addition of Mg masses below 7 mg did not significantly change LHGA with DBC 0.060±0.004 0.172±0.060 LHGA without DBC 0.226±0.003 — THGA with ZBC 0.040±0.001 0.004±0.000 THGA without ZBG 0.040±0.001 — aMean±one standard deviation of 10 readings. No chemical modifier was used. Experimental conditions as specified in Table 1. important to note that the uncorrected (A s+BG) measurements were not as precise as the corrected signals, since some sample components that cannot be driven oV during pyrolysis cause BG absorption.This eVect could be even more noticeable when samples from subjects with diVerent ailments, which may contain diVerent amounts of organic components, are analyzed. Hence, it was considered that further research should be carried out using an appropriate chemical modifier that would stabilize the analyte during the pre-atomization steps, with the subsequent reduction of the higher BG signals, using the most sensitive approach, i.e., using PA atomization.Choice of chemical modifier and atomizer with diVerent heating modes The operating parameters chosen for chromium determination Fig. 2 Atomization signals for 0.03 ng of Cr in urine samples without (Table 1) were selected to achieve maximum sensitivity and and with chemical modifier using platform atomization in LHGA and precision and lower background signals. The graphite furnace DBC.(a) Without modifier; (b) with 0.8 ng of Ni; (c) with 0.8 ng of settings for both instruments used were determined experimen- Pd; (d) with 8.2 mg of Mg; (e) with 0.4 ng of Eu; and (f ) with 0.4 ng of Eu and 0.8 ng of Pd. Other conditions as specified in Table 1. tally by varying the drying, pyrolysis and atomization times J. Anal. At. Spectrom., 1999, 14, 821–825 823Fig. 4 Pyrolysis and atomization temperatures. (A) Using LHGA Fig. 3 Atomization signals for 0.08 ng of Cr in aqueous solutions and with DBC for 0.08 ng of Cr, (a) with 0.4 ng of Eu as modifier and (b) 0.03 ng of Cr in urine samples using platform atomization in THGA without the addition of a modifier, and for a urine sample, (c) with and ZBC without the addition of a chemical modifier and adding Mg. 0.4 ng of Eu as modifier and (d) without a modifier. (B) (a)–(d) (A) Aqueous solution, (a) without and (b) with 8.2 mg of Mg as correspond to the integrated absorbance signals obtained using THGA modifier using the Tpyr; (B) urine sample without modifier using the with ZBC (in this case Mg 8.2 mg was added as chemical modifier).Tpyr; (C) urine sample without Mg using a pyrolysis temperature of Other conditions as specified in Table 1. 1200 °C; and (D) urine sample with 8.2 mg of Mg using the Tpyr. Other conditions as specified in Table 1. the sensitivity for the determination of chromium. Sensitivity with their respective BG correction procedures are shown in Fig. 4. It should be noted that there is a variation between the losses with increasing modifier masses above 10 mg were set and real atomizer pyrolysis and drying temperatures observed regardless of the atomization temperature or long depending on the tube conditions, in particular the electrical integration times. Based on the previously described results, resistance between the tube and contacting cones. Therefore, further research was only conducted by adding Eu and Mg experimental parameters that allow for temperature variations for LHGA with DBC and THGA with ZBC, respectively.should be selected. It is critical to choose Tpyr as was done in this work. Although when the DBC was used the variation of Temperature program the pyrolysis and atomization temperatures were diVerent in Complete dryness of aqueous standards and samples was the presence of Eu, its addition stabilized the chromium signals ensured with two drying steps; 10 and 5 s ramps for the second in both matrices at a pyrolysis temperature of 1500 °C, which drying temperature at 120 and 130 °C avoided spattering of was subsequently considered as the Tpyr.When the ZBC was the liquids and resulted in a uniform solid deposit on the used the addition of Mg stabilized the integrated signals up to surface of the atomizer platforms of the PE 2100 and PE 4100 a temperature of 1400 °C, with a considerable decrease subinstruments, respectively. sequently.This temperature was considered as the Tpyr, The pyrolysis and atomization curves for chromium aqueous because at lower pyrolysis temperatures too high BG signals were observed (Fig. 3C). solutions and urine samples using both types of atomizers Table 3 Analytical characteristics Regression Linear range/ Detection limit/ Instrument Modifier equationa mg l-1 r2 RSD (%) mg l-1 mo/pg PE 2100 0.4 ng of Eu A=0.0064+0.0399 [Cr] 0–20 0.9988 2.0–2.5 0.02b (0.4 pg) 2.2 0.03c (0.6 pg) PE 4100 8.2 mg ofMg A=0.0071+0.0267 [Cr] 0–10 0.9992 0.8–1.4 0.04b (0.8 pg) 3.3 0.05c (1.0 pg) a[Cr]=chromium concentration in mg l-1. bDetection limits for aqueous solutions.cDetection limits for a urine sample with low chromium concentration. Table 4 Determination of chromium in certified urine samples and samples from non-environmentally exposed subjects and stainless-steel factory workers Cr in the subjects under study/mg l-1 Cr in certified samplesa/mg l-1 Non-exposed Exposed Instrument Seronorm NIST (n=20) (n=41) PE 2100 with 20.0±1.0 0.087±0.003 0.14±0.09 0.47±0.22 LHGA and DBC PE 4100 with 20.0±0.8 0.086±0.002 0.14±0.08 0.46±0.27 THGA and ZBC aRecommended [Cr]=20±1 mg l-1 in Seronorm Trace Elements in Urine and 0.085 mg l-1 in SRM 2670 Freeze-dried Urine from NIST. 824 J. Anal. At. Spectrom., 1999, 14, 821–825Analytical performance posed subjects and stainless-steel factory workers using both types of atomizers are in good agreement and have good Calibration curves were prepared with aqueous chromium precision.Depending on instrument availability, any of the standards following the procedure described above and the compared procedures can be applied to the determination of temperature program given in Table 1. The linear ranges, the chromium in urine of ‘normal’ subjects for baseline studies or limits of detections, characteristic masses and precision [calcu- to evaluate environmental or other non-occupational exposure lated as relative standard deviation (RSD) for six consecutive to chromium. measurements] are given in Table 3.The within-batch RSD were found to be 2.0–3.5 and 0.8–1.4% with the LHGA and Acknowledgements THGA, respectively. They also show the improved sensitivity (slope of the calibration graphs) obtained with the LHGA, The authors are grateful to the CDCHT (Consejo de probably because with these atomizers the rate of diVusional Desarrollo Cientý�fico, Humaný�stico y Tecnolo� gico) of the loss of analyte is reduced.23 The slopes of the calibration Andes University for financial support.graphs of chromium in urine were not statistically distinguishable from those of aqueous standards presented in Table 3. References The limit of detection (LOD), defined as three times the standard deviation of a blank solution, was 0.4 and 0.8 pg for 1 R. A. Goyer, in Casarett and Doull’s Toxicology: the Basic Science aqueous chromium standards using the longitudinally and of Poisons, ed.C. D. Klassen, M. O. Amdur and J. Doull, Macmillan, New York, 3rd edn., 1986. transversally heated atomizers, respectively. The LOD for the 2 E. J. Underwood, Trace Elements in Human and Animal Nutrition, method was also calculated by repeated analysis of a urine Academic Press, New York, 4th edn., 1977. sample with a low chromium concentration. In this case, the 3 W. Slavin, G. R. Carnrick, D. C. Manning and E. Pruszkowska, LODs were 0.03 and 0.05 mg l-1 using the longitudinally and At. Spectrosc., 1983, 4, 69.transversally heated atomizers, respectively, which were higher 4 K. Ohta, H. R. Uegomori, S. Itoh and T. Mizuno, Microchem. J., than those obtained for aqueous solutions owing to the 1997, 56, 343. 5 E. Alvarez-Cabal Cimadevilla, K. Wro� bel, J. M. Marchante presence of some matrix components. The results for the Gayo�n and A. Sanz-Medel, J. Anal. At. Spectrom., 1994, 9, 117. determination of chromium in the SRM are summarized in 6 A.Taylor, S. Branch, D. J. Halls, L. M. W. Owen and M. White, Table 4. Good agreement was obtained using either BG correc- J. Anal. At. Spectrom., 1998, 13, 67R. tion procedure. 7 B. L. Gong and Y. M. Liu, At. Spectrosc., 1990, 11, 229. The mo using both BG correction types found in this study 8 S. T. SauerhoV, Z. A. Grosser and G. R. Carnrick, At. Spectrosc., was lower than those previously reported by Paschal and 1996, 17, 273. 9 R.Rubio, A. Sahuquillo, G. Rauret, L. Garcia Beltran and Bailey12 (mo=3.0 pg) using nitric acid+Triton X-100 as the Ph. Quevauviller, Anal. Chim. Acta, 1993, 283, 207. chemical modifier, Thoml.26 using Mg (mo=3.3 pg), 10 J. N. Marks, M. A. White and A. R. Boran, At. Spectrosc., 1988, Rh (mo=3.0 pg) and Pt (mo=2.8 pg) as chemical modifiers, 9, 73. Granadillo et al.15 using a fast furnace program without the 11 J. Kumpulainen, J. Lehto, P. Koivvistoinen, M. Uusitupa and addition of any chemical modifier (but treating the sample E.Vuori, Sci. Total Environ., 1983, 31, 71. with HN03) (mo=2.7 and 5.0 pg using Perkin-Elmer Model 12 D. C. Paschal and G. G. Bailey, At. Spectrosc., 1991, 12, 151. 13 P. Chappuis, J. Poupon, J. F. Deschamps, P. J. Guillausseau and HGA and 5100 PC instruments, respectively), Alvarez-Cabal F. Rousselet, Biol. Trace Elem. Res., 1992, 32, 85. Cimadevilla et al.5 using wall atomization in the absence of a 14 C. Veillon, K.Y. Patterson and N. A. Bryden, Clin. Chem., 1982, chemical modifier and Slavin and Carnrick27 using MgNO3 28, 2309. (mo=3.3 pg) as chemical modifier and ZBC. Hence the most 15 V. A. Granadillo, L. P. de Machado and R. A. Romero, Anal. significant improvement made in this work for the determi- Chem., 1994, 66, 3624.. nation of Cr in urine samples is the increase of sensitivity, 16 W. Slavin, Graphite Furnace AAS. A Source Book, Perkin-Elmer, Norwalk, CT, 1984. which includes the possibility of using DBC but in the presence 17 O.Gene, S. Akman, A. R. Ozdural, S. Ates and T. Balkis, of an appropriate chemical modifier such as Eu. Spectrochim. Acta, Part B, 1981, 36, 163. 18 J. L. Burguera, M. Burguera, C. E. Rondo�n and E. Burguera, At. Analytical application Spectrosc., 1997, 18, 109. 19 P. E. Burguera, A. Sanchez de Bricen� o, C. E. Rondo� n, J. L. Chromium in urine samples from 20 non-environmentally or Burguera, M. Burguera and P. Carrero, J.Trace Elem. Med. Biol., occupationally exposed subjects and from 41 workers in a 1998, 12, 115. stainless-steel factory was determined using both types of 20 M. Burguera, J. L. Burguera, C. Rondo� n, M. L. di Bernardo, M. Gallignani, E. Nieto and J. Salinas, Spectrochim. Acta, Part B, atomizer with their respective BG correction procedures. The 1999, in the press. results are given in Table 4. Although the chromium levels 21 M. Burguera, J. L. Burguera, C. Rondo�n, M. R. Brunetto and were statistically significantly higher ( p<0.01) in the exposed C. Rivas, in Metal Ions in Biology and Medicine, ed. Ph. Collery, population, in both cases the metal content may be considered P. Bratter, V. Negretti de Bratter, L. Khassanova and J. C. to be within ‘normal’ chromium levels (below 1.0 mg l-1).28,29 Etienne, Libbey Eurotext, Paris, 1998, vol. 5, pp. 13–17. Good agreement with the previous results was obtained for 22 G. Schlemmer, H. Schulze and C. Gu� nner, Analytical Advantages of End-capped Tubes Used with a Transverse-heated Graphite Cr for both BG correction procedures with a linear regression Atomizer, Technical Summary Order No. TSAA-46(B050–4288), equation y=1.002x-0.050 (r=0.998), where x and y are Perkin-Elmer, U� berlingen, 1995. chromium concentrations in mg l-1. This indicates that the 23 J. M. Harnly and B. Radziuk, J. Anal. At. Spectrom., 1995, 10, two BG correction procedures did not give significantly diVer- 197. ent values. 24 A. Taylor and P. Green, J. Anal At. Spectrom., 1988, 3, 115. 25 D. J. Halls and G. S. Fell, J. Anal. At. Spectrom., 1986, 1, 135. 26 N. S. Thomaidis, E. A. Piperaki, C. R. Polydorou and C. E. Conclusion Efstathiou, J. Anal. At. Spectrom., 1996, 11, 31. 27 W. Slavin and G. R. Carnrick, At. Spectrosc., 1985, 6, 157. In the light of this comparative study, it is concluded that Eu 28 J. Versieck and R. Cornelis, Trace Elemments in Human Plasma or and Mg (as magnesium nitrate) provide the best results, in Serum, CRC Press, Boca Raton, FL, 1989. terms of thermal stabilization and sensitivity, when LHGA 29 K. S. Subramanian, Prog. Anal. Chem., 1988, 11, 513. with DBC and THGA with ZBG are used, respectively. The results obtained for urine samples from environmentally unex- Paper 8/08239A J. Anal. At. Spectrom., 1999, 14, 821

ISSN:0267-9477    

DOI:10.1039/a808239a

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Characterization and vapour phase interference studies of a flame heated holed quartz T-tube as atomization cell for hydride generation atomic absorption spectrometry† Patricia Grinberg,*a Iracema Takaseb and Reinaldo Calixto de Camposa aDepartment of Chemistry, Pontifý� cia Universidade Cato�lico do Rio de Janeiro, Rio de Janeiro, Brazil. E-mail: grinberg@rdc.puc-rio.br, rccampos@rdc.puc-rio.br bDepartment of Analytical Chemistry, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil.E-mail: takaseir@iq.ufrj.br Received 23rd November 1998, Accepted 24th March 1999 Characterization and mutual hydride forming element interference studies of a flame heated holed quartz T-tube as atomization cell for HGAAS were performed. Characteristic concentrations, limits of detection and linear ranges were determined for As, Bi, Sb and Se. Also, the mutual interferences of these elements were investigated. Extended non-interferent ranges were observed in most of the studied cases and such improvement was obtained at no great expense with regard to sensitivity and detection limits. Mutual hydride forming element interference is a well known Procedure phenomenon in hydride generation atomic absorption spec- All plastic and glassware were immersed in 10% HNO3 for at trometry (HGAAS).1–10 According to the current literaleast 24 h and thoroughly rinsed with Milli-Q water before ture,11,12 the atomization mechanism in HGAAS involves a use.The reaction flask was washed with Milli-Q water between H radical mediated process and the competition for this radical successive uses to avoid memory eVects and contamination.explains these interferences. Indeed, it has been verified that, The T-tubes were pre-heated for at least 10 min before the in environments richer in H radicals, larger sensitivities may first measurement and the reagent blank value was tested be reached and mutual hydride forming element interferences periodically and always subtracted from the measured analyte are minimized.5,13 The aim of the present work is to investigate signals.All measurements were performed using 10 mL of 1, a flame heated holed quartz T-tube as an alternative atomiz- 5 and 10% v/v HCl in the reaction flask, in peak height, and ation cell for HGAAS, with special regard to mutual hydride all data given in this work are the average of five independent forming element interferences. determinations.Experimental Results Instrumentation Sensitivities, instrumental detection limits and linear ranges A Perkin-Elmer Model (Norwalk, CT, USA) MHS-10 hydride Table 2 compares the figures of merit for both tested atomizers generation batch system was used. Nitrogen was used as purge using a stoichiometric flame. Sensitivities (c0, characteristic and carrier gas. The generated hydrides were atomized in two types of flame heated quartz cell, diVering from each other in the presence or absence of holes (Fig. 1). All measurements were performed with a Varian Techtron (Palo Alto, CA, USA) Model AA5 atomic absorption spectrometer using an air– acetylene flame. Perkin-Elmer hollow cathode lamps (HCL) were used for all elements. The operating parameters and the instrumental settings were adjusted according to the manufacturers’ recommendations and some of them are listed in Table 1. No background correction was used. Reagents and Solutions Milli-Q water (Millipore-Waters, Milford, MA, USA) was used throughout and all chemicals were of analytical-reagent grade; 1.5% m/v NaBH4 (Vetec) was freshly prepared by od 2 id 10, od 12 12 51 96 18 18 17 18 dissolving the salt in 0.5% m/v NaOH (Vetec) and this solution Fig. 1 Holed quartz T-tube. All measurements are in millimetres. was always filtered before use. Analytical solutions, as well as Table 1 Operating parameters those containing interferents, were prepared from convenient dilutions and spikings of Tritisol concentrates (1000 mg L-1, Parameter As Bi Sb Se Merck, Elmsford, NY, USA) to the reaction flask. Wavelength/nm 193.7 223.1 217.6 196 Slit/nm 0.3 0.05 0.1 0.3 †Presented at the Fifth Rio Symposium on Atomic Spectrometry, HCL current/mA 7 8 10 10 Cancu� n, Mexico, October 4–10, 1998.J. Anal. At. Spectrom., 1999, 14, 827–830 827Table 2 Sensitivities (c0), detection limits and linear ranges (all in mg L-1) in the determination of As, Bi, Sb and Se in 10% v/v HCl by HGAAS using a conventional atomizer and the proposed atomizer Quartz T-tube (QTA) Holed quartz T-tube (HQTA) Characteristic Detection Linear range Characteristic Detection Linear range Element concentration limit (up to) concentration limit (up to) As 0.11 0.06 15 0.47 0.23 40 Bi 0.28 0.15 30 0.39 0.20 60 Sb 0.31 0.17 35 0.36 0.18 35 Se 0.17 0.09 30 0.25 0.16 40 Fig. 2 Interferences of Bi (+), Sb (#), Se (>) and Te (%) on the As (5 mg L-1) signal in 10 mL of 10% HCl: (a) conventional quartz T-tube; (b) holed quartz T-tube.Fig. 3 Interferences of As ($), Sb (#), Se (>) and Te (%) on the Bi (5 mg L-1) signal in 10 mL of 10% HCl: (a) conventional quartz T-tube; (b) holed quartz T-tube. concentration) and linear ranges (10% deviation from the No significant change in these analytical figures was observed in relation to the other HCl concentrations (1 and 5%) studied. linear projection) were taken from aqueous analytical curves. The instrumental detection limits were calculated from LOD= Table 2 shows that, for both atomizers, comparable characteristic concentrations and detection limits were 3ss/m, where ss is the estimated standard deviation of ten blank measurements and m is the slope of the analytical curve.observed for all tested elements, except As. Concerning the 828 J. Anal. At. Spectrom., 1999, 14, 827–830Fig. 4 Interferences of As ($), Bi (+), Se (>) and Te (%) on the Sb (20 mg L-1) signal in 10 mL of 10% HCl: (a) conventional quartz T-tube; (b) holed quartz T-tube.Fig. 5 Interferences of As ($), Bi (+), Sb (#) and Te (%) on the Se (20 mg L-1) signal in 10 mL of 10% HCl: (a) conventional quartz T-tube; (b) holed quartz T-tube. Table 3 Tolerance limits (±10%) for the studied interferent elements using a conventional atomizer and the proposed atomizer Analyte As (5 mg L-1) Bi(5mg L-1) Sb(20mg L-1) Se(20mg L-1) Interferent QTA HQTA QTA HQTA QTA HQTA QTA HQTA As — — >104 >104 >104 >104 1000 5000 Bi >104 >104 — — 7000 5000 2000 2000 Sb 300 5000 1000 >104 — — 500 >104 Se 50 2000 500 >104 300 5000 — — Te 5000 >104 600 3000 >104 >104 >104 >104 J.Anal. At. Spectrom., 1999, 14, 827–830 829linear ranges (upper concentration/LOD), extended linear concentration range. Such improvement was obtained at no large expense with regard to sensitivity and detection limits, ranges using the proposed atomizer (HQTA) were observed for all elements, except Sb.except for As where a fivefold decrease in sensitivity was observed. However, an extended linear range was observed in most cases. The use of a lean flame was also tried, with much Mutual hydride forming element interferences better tolerance limits for the HQTA, but with a large sensi- The interference behaviour for both tubes in 10% v/v HCl is tivity drop. Thus, the use of this flame is not advisable. displayed in Figs. 2–5.The relative signal is referred to as the ratio between the absorbance in the presence of the interferent Acknowledgements and that of the pure analyte solution. A similar behaviour was observed for the other HCl concentrations (1 and 5%) studied. The Brazilian National Research Council is acknowledged for The tolerance limits (±10% change in the absorption signal ) the provision of financial support. in 10% v/v HCl for the four hydride forming elements studied are summarized in Table 3.References Among the studied elements, Sb and Se were the most critical interferents for As. However, the use of the proposed 1 P. Barth, V. Krivan and R. Hausbeck, Anal. Chim. Acta, 1992, 263, 111. atomizer extended significantly the non-interferent range in 2 B. Welz and M. Melcher, Anal. Chim. Acta, 1981, 131, 17. both cases. In the case of Te as intnt, an improvement 3 K. Dittrich and R. Mandry, Analyst, 1986, 11, 269. of at least twofold was observed and, for Bi, the atomizers 4 M.Yamamoto, M. Yasuda and Y. Yamamoto, Anal. Chem., 1985, were comparable. For Bi as analyte, the proposed atomizer 57, 1382. was also able to extend the non-interferent range for all cases 5 J. Dedina, Anal. Chem., 1982, 54, 2097. of interference. Sb and Se (as analytes) were the only two 6 L. Lajunen, T. Merkkiniemi and H. Hayyrynen, Talanta, 1984, 31(9), 709. cases where no improvement was observed, both related to 7 G. Hall and J. Pelchat, J. Anal. At. Spectrom., 1997, 12, 97. the Bi interference. However, in all other situations in which 8 M. B. de la Calle, R. Torralba, Y. Madrid and W. Palacios, an interference eVect was noticed, the tolerance limit was Spectrochim. Acta, Part B, 1992, 47(10), 1165. increased by the HQTA. 9 J. Hershey and P. Keliher, Spectrochim. Acta, Part B, 1986, 41(7), 713. 10 K. Dittrich, R. Mandry and U. Udelnow, Fresenius’ Z. Anal. Final discussion and conclusion Chem., 1986, 323, 793. 11 J. Dedina and I. Rubeska, Spectrochim. Acta, Part B, 1980, 35, The holed quartz T-tube was able to induce larger tolerance 119. limits of interference in nine of the 16 mutual interference 12 B. Welz and M. Melcher, Analyst, 1983, 108, 213. possibilities studied and a comparable performance with the 13 J. Dedina, Spectrochim. Acta, Part B, 1992, 47, 689. conventional T-tube was achieved in two cases. For the five remaining cases, no interference was observed in the studied Paper 8/09154D 830 J. Anal. At. Spectrom., 1999, 14, 827–830

ISSN:0267-9477    

DOI:10.1039/a809154d

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

On the use of line intensity ratios and power adjustments to control matrix eVects in inductively coupled plasma optical emission spectrometry E. H. van Veen* and M. T. C. de Loos-Vollebregt Laboratory of Materials Science, Delft University of Technology, Rotterdamseweg 137, 2628 AL Delft, The Netherlands, E-mail: eric.vanveen@stm.tudelft.nl Received 14th October 1998, Accepted 16th February 1999 In inductively coupled plasma optical emission spectrometry, matrix eVects can be substantially reduced by applying robust operating conditions, i.e.a high rf power level and a low nebulizer gas flow. However, dissimilar line intensity changes are still observed, in particular with varying salt matrices. Calcium and, to a lesser extent, Mg induce stronger eVects than Na and K. In 0.1, 0.3 and 1.0% Ca matrices, signal changes for 29 atomic as well as ionic lines have been determined with respect to the Ca-free solution. The changes range from -5 to -30% for the 1.0% Ca matrix.Starting from robust conditions in radial viewing, the rf power has been adjusted until the Cr ionto- atom line ratio measured for the calcium solutions equalled the ratio determined during the calibration (0% Ca). By this power adjustment, the changes are within ±4%. No matrix eVects originating from the sample introduction system are observed, probably due to the high acid concentrations used. The practical application of power adjustments is illustrated with results for certified sediment samples and with multiple line analysis for qualitative and semiquantitative analysis. The approach is an attractive alternative to matrix matching or standard additions.Internal standardization based on one atomic and one ionic line of the same element is indicated as another possibility. When analysing series of samples in inductively coupled plasma The Mg II/I line ratio has also been used as a fast diagnostic optical emission spectrometry (ICP-OES), the sample matrices tool in the control of ICP systems,10 and in qualitative and may vary in acid and/or salt composition.Variations in acid semiquantitative analysis.11–13 In this kind of analysis, the composition are not expected to be large, as fixed amounts assessment of the presence of an element is based on the are usually applied for digestion or storage of samples. The detection of the prominent emission lines and their relative main components may vary appreciably, such as the transition line intensities compared with predetermined values of the elements in steel or the alkali and alkaline earth elements in pure element solution.To allow the comparison of relative sediment digestions. For proper calibration, one has to con- line intensities, the plasma conditions on analysis should be sider the sample matrix composition, and methods such as close to the conditions at which the reference values were matrix matching, internal standardization, standard addition, determined.In verification, the Mg II/I line ratio should be optimization of ICP operating conditions and mathematical reproduced at its high value indicating robust and identical correction can be applied.1 With matrix matching, the known conditions. Otherwise, it was suggested12 that the rf power amount of acid and the expected average amount of salt (if level should be adjusted slightly in order to reproduce the Mg necessary) are added to all calibration solutions.II/I line ratio. According to the insights obtained by Mermet’s group,2–9 Present-day ICP-OES instrumentation14–17 includes an matrix eVects have their origin in diVerent parts of the ICP echelle grating in combination with a cross disperser to obtain instrument, viz. in the plasma itself and in the sample introduc- a two-dimensional image of the emission spectrum. Charge tion system. It has been shown that the plasma can be made transfer device detectors are then applied to generate the robust with respect to atomization, ionization and excitation digital representation of the full spectrum.This instrumenby appropriate selection of the operating conditions.2–4 The tation has allowed the development of semiquantitative survey matrix eVects can be substantially reduced by using operating analysis, where the full sample spectrum is compared with a conditions that lead to an eYcient energy transfer between the set of pure element spectra.18,19 Based on multicomponent plasma and the sample.With a high rf power level (1200 W), analysis, all elements and all of their lines present in the a low nebulizer gas flow (0.6 l min-1) and a wide injector spectrum are processed simultaneously. As about 70 elements inner diameter (2 mm), all ionic lines having an energy sum can be detected by ICP-OES, the set of spectra should be between 8 and 16 eV are suppressed by more or less the same determined in a once-only calibration using pure element amount in the presence of a matrix as compared to the matrixsolutions. The solutions contain 2% HNO3 as the common free situation.The suppression is quite insensitive to significant matrix and the spectra include a large amount of atomic and changes in the salt or acid matrix composition, and may ionic lines with widely diVerent energy sums. This predeter- originate from changes in the aerosol transport in the sample mined set of spectra is intended not only to calibrate sample introduction system.5–7 It can be compensated for by using signals over the lifetime of the instrument, but also to calibrate matrix matched standards or internal standardization.8 Robust sample signals in divergent matrices.In order to reproduce plasma conditions are characterized by the high intensity ratio the line patterns at a later time or in a diVerent matrix (HCl, (>8) of the Mg II 280.270 nm line with respect to the Mg I Na, K and Ca), rf power adjustments have successfully been 285.213 nm line.9 Hence, the Mg II/I line ratio for standard applied19 using a matrix matched monitor solution containing and sample solutions can be measured to confirm the robustness of the analysis with respect to matrix eVects.Ba, Cd, Cr, Cu and Ga and employing the line ratio of the J. Anal. At. Spectrom., 1999, 14, 831–838 831Table 1 Operating conditions of the Perkin-Elmer Optima 3000 DV Cr II 267.716 nm and Cr I 357.869 nm lines as the sensitive spectrometer criterion for appropriate adjustment.Thompson et al.20 found that matrix-induced changes in Nominal rf power/W 1300 excitation conditions in the plasma and a change in the rf Plasma gas flow/l min-1 15 power supplied to the plasma produced a similar eVect. This Auxiliary gas flow/l min-1 0.5 Nebulizer gas flow/l min-1 0.6 similarity was used to compensate for matrix eVects as large Sample uptake rate/ml min-1 1.0 as 30%.To all test solutions with Al, Ca, K, Mg, P and S as Nebulizer Cross-flow the matrix elements, the same Be amount was added and the Diameter of injector tube/mm 2.0 rf power was automatically tuned proportional to the diVerence Viewing Radial between the observed and the reference Be II intensity at View distance from coil/mm 4 313 nm. Budic¡ and Hudnik21 also increased the power level Resolution Normal Read time Auto by a small amount to correct for KCl and H3PO4 matrix Minimum and maximum read time/s 2 eVects which were found to correlate with the excitation energy of the ionic lines. Mermet7 reported that, when applying robust conditions, matrix eVects can be minimized to almost the same extent, ing 5% HNO3 and 0, 0.1, 0.3 or 1.0% Ca. Suprapure regardless of the elements, line characteristics and radial or CaCl2.4H2O was used.axial viewing mode. However, close inspection of the available The sediments BCR145, BCR277, BCR280, BCR320 and data4,6,22 shows that the signal suppression still varies over Maas were digested by dissolving about 2 g in 100 ml of 16% the analytes or, even worse, over diVerent ion lines of the aqua regia.The Maas sediment is distributed by the Institute for same analyte. In the work of Brenner et al.,23 matrix eVects InlandWaterManagement andWastewater Treatment, Lelystad, owing to 1 g l-1 Ca or Na were observed to be relatively The Netherlands. Calibration was performed using 10 mg l-1 small, but suppressions for most of the analytes were dissimilar.Merck ICP multielement standard IV in 16% aqua regia. Such changes with matrix composition point at changes in plasma conditions. The main objective of this paper is to Results and discussion further investigate the dissimilar intensity changes of emission lines in ICP spectra. Starting from robust conditions, it will Robust conditions be shown to what extent line intensities can be reproduced, how rf power adjustments can help and how ion-to-atom line Robust plasma conditions are applied as the reference conditions in the present study.As in the work of Mermet’s ratios can serve as the criterion. As solutions containing Ca exhibit a large matrix eVect as compared to solutions contain- group, the nebulizer flow was set as low as 0.6 l min-1 and the standard alumina injector tube was mounted. Among the ing Na, K or Mg, the Ca matrix is used as the worst case example. The approach of power adjustments is illustrated available injectors, this tube has the largest inner diameter of 2.0 mm.The rf power was set to 1300 W. Although a more with several certified sediment samples. Herewith, a simple solution to an old and persistent problem of Zn analysis in robust plasma can be obtained at the software controlled maximum power level of 1500 W, the lower setting was used sediments is presented. In addition, it is shown how semiquantitative, multiple line analysis can benefit from power to allow power adjustments when nebulizing heavy matrices.The optimum distance from the rf coil was determined by adjustments. nebulizing a test solution containing Mg and Cr. Intensities were measured for the Mg 280.270 nm and the Cr 267.716 nm ion lines and for the Mg 285.213 nm and the Cr 357.869 nm Experimental atom lines. The normalized line intensities as a function of ICP emission spectrometer and settings distance are displayed in Fig. 1. The atom lines attain their maximum intensity around 4 mm, whereas the ion line intensit- This study was carried out with a Perkin-Elmer (Norwalk, ies continue to increase at smaller distances.Fig. 2 shows the CT, USA) Optima 3000 DV system. The sample introduction normalized ion-to-atom line ratios for Mg and Cr. In order system consisted of a cross-flow nebulizer and a double-pass not to measure too close to the coil, the 4 mm setting was Scott-type spray chamber. According to the procedure selected as the proper distance.This optimum distance of a described below, robust plasma conditions were realized by few millimetres from the coil was also selected in other work.3,4 applying 1300 W rf power and 0.6 l min-1 nebulizer gas flow. As can be seen in Fig. 1, the Mg and Cr ion line intensities The injector inner diameter was 2.0 mm. Most data were show the same behaviour as a function of distance from the obtained in the radial viewing mode at 4 mm distance from coil.The Mg and Cr atom line intensities, however, behave the load coil. diVerently: there is almost no change in the Cr I intensity. The spectrometer consists of an echelle grating and separate This is reflected in the ion-to-atom line ratios in Fig. 2: the Cr cross-dispersers for the UV and visible channels. Both optical II/I ratio has the largest relative increase with decreasing channels end up in separate segmented charge coupled device distance. The same is true for the Cr II/I ratio with increasing (SCD) detectors.The system measures simultaneously 6% of rf power level, as displayed in Fig. 3 together with a large set the continuous ICP spectrum from 167 to 782 nm on 201 of relative line ratios. This example was determined for a 0.1% subarrays. The subarrays cover one to four prominent Ca matrix relative to the matrix-free situation (at 1300 W). analytical lines for each detectable element. By using the multielement standard and subarrays available All pertinent operating conditions are summarized in on the Perkin-Elmer Optima 3000 DV system, the change in Table 1. the relative line ratios as a function of rf power was determined for the combinations: Cd 214/228, Cd 226/228, Cr 205/357, Standards and samples Cr 267/357, Cu 224/324, Cu 224/327, In 230/325, Mg 280/285, Ni 221/232, Ni 231/232, Pb 220/283, Tl 190/276, Zn 202/213 Test solutions of 1 mg l-1 Mg and 20 mg l-1 Cr were prepared in 5% HNO3 and in 1000 mg l-1 Na, K, Mg or Ca using and Zn 206/213. From Fig. 3, the change is similarly large for the Cr 205/357, Cr 267/357 and Pb 220/283 combinations, Merck (Darmstadt, Germany) single element standards. To study the Ca matrix eVect, 10 mg l-1 dilutions of the Merck smallest for the Mg 280/285, Ni 231/232 and Tl 190/276 combinations, and in between for the two Zn combinations. ICP multielement standard IV were made in solutions contain- 832 J. Anal. At. Spectrom., 1999, 14, 831–838Fig. 1 Ionic and atomic line intensities for Mg and Cr as a function of the distance from the rf coil, normalized to the intensities at 1 mm distance. Rf power is 1300 W. 1, Mg II 280.270 nm; %, Mg I 285.213 nm; ×, Cr II 267.716 nm; *, Cr I 357.869 nm. Fig. 2 Ion-to-atom line ratios for Mg and Cr as a function of the distance from the rf coil, normalized to the ratios at 1 mm distance. Rf power is 1300 W. 1, Mg II/I; ×, Cr II/I. In an earlier study,19 the Cr 267/357 ratio was found to DV instrument and by making the reasonable assumption that the ICP continuum background intensities are the same over monitor, on average, the eYciency of atomization and ionization for the full set of elements detectable in ICP-OES. Owing a wavelength range of only 5 nm, our correction value equals 1.34.This results in an Mg II/I ratio of 10.7, confirming that to this fact and to its strong dependence on the rf power, the Cr 267/357 line ratio will be used in the present study when the ICP operates under robust conditions.It is clear that robustness can also be characterized using correcting matrix eVects by power adjustments. Robustness, nevertheless, has been characterized in general the Cr II/I ratio. However, eYciency corrections cannot easily be made in order to report values which can be compared by the Mg II/I ratio, which should be larger than 8.7 As in the Optima 3000 DV system the actual rf power applied to from instrument to instrument, because of the large diVerence between the wavelengths of the two chromium lines.the ICP can diVer from the nominal setting by ±50 W, the ratio was determined over several days for the operating conditions given in Table 1. The obtained average of 14.4 Matrix eVects from Na, K, Mg and Ca should be corrected for diVerences in the eYciencies of the echelle grating and the SCD detector. For one specific Optima Under robust conditions, changes in the plasma conditions are expected to be small when the matrix concentration is 3000 DV instrument, a correction factor of 1.85 was reported.24 Although the authors did not explain how they derived this significantly modified. With an acidic matrix, changes in the amount of acid cause only minor changes in ionic line intensit- value, the correction factor has been used by others for their instruments.6,25–27 By measuring the background intensities in ies.2 Elements such as Na, K, Mg and Ca, however, are notorious for their matrix eVects and, even under robust the Mg 280 nm and Mg 285 nm subarrays of our Optima 3000 J.Anal. At. Spectrom., 1999, 14, 831–838 833Table 3 Lines (nm) used in the determination of line intensity ratios Al I 237 Cr II 205 Ga I 294 Pb II 220 Al I 396 Cr II 267 Ga I 417 Pb I 283 Ba II 455 Cr I 357 In II 230 Zn II 202 Cd II 214 Cu II 224 In I 325 Zn II 206 Cd II 226 Cu I 324 Mn II 257 Zn I 213 Cd I 228 Cu I 327 Ni II 221 Co II 228 Fe II 238 Ni II 231 Co II 230 Fe II 259 Ni I 232 including 5% HNO3 at 1300 W.Fig. 4 shows the line intensity ratios with and without the presence of 1% Ca as a function of the energy sum of the 29 lines. For the present Optima 3000 DV instrument, the observed change in the ratio is more or less linear with the energy sum and suggests that a change should be made to one of the plasma parameters. Here, we will adjust the rf power level to investigate to what extent the eVects owing to diVerent Ca matrices can be corrected for with the Cr II/I ratio as the criterion.Although the feedback electronics define the ultimate power level when switching on Fig. 3 Ion-to-atom line ratios with and without the presence of 0.1% the plasma (±50 W with respect to the nominal value), the Ca as a function of the rf power level. Distance from coil is 4 mm. power level is stable after the warming-up time and can be The dotted line indicates a relative line ratio equal to 1.tuned to any desired level in a reproducible way.18 The ratios of the line intensities measured in the matrix and conditions, dissimilar suppressive eVects for various ionic lines in 5% HNO3 vs. the rf power are displayed in Fig. 5(a) for up to tens of per cent have been observed.4,6,22,23 As a the 0.1% Ca matrix and in Fig. 5(b) for the 1.0% Ca matrix. consequence, the suppressive eVect may not (only) be due to In Fig. 5(a), it can be seen that, at 1300 W, some lines are the sample introduction system,5–7 but may be due to some suppressed by 10% while other lines are not suppressed at all.further change in plasma conditions. By increasing the power, the spread in suppression decreases, In order to clearly observe the remaining matrix eVect under attains a minimum, the suppression changes into an enhancerobust conditions, the alkali and alkaline earth elements were ment and the spread in enhancement increases. The minimum investigated to see which produced the strongest eVect.Mg spread occurs at the rf power level at which the line intensity and Cr line intensities were measured in solutions containing ratios for the Cr II 267 nm and Cr I 357 nm lines are equal. 1000 mg l-1 of Na, K, Mg or Ca. Table 2 shows that the Or, in other words, at the optimum power setting, the Cr II/I plasma is fully robust with respect to the Na and K matrices. ratio in the matrix equals the reference value. The same For the Mg and Ca matrices, the Cr ion line is suppressed, behaviour is observed for the two other matrices, where, with whereas the atom line is not, and the Cr II/I ratio indicates a increasing amount of Ca, the initial spread in the suppressions change in the plasma conditions.The Ca matrix shows the and the suppressions themselves increase and the optimum largest eVect, as also observed by others,19,23 and was selected power level occurs at a higher setting, as is shown for the to further investigate the dissimilar line intensity changes. 1.0% Ca matrix in Fig. 5(b). The Mg intensities varied over the respective matrices, owing Table 4 lists the averaged suppression, i.e. the matrix eVect, to the small contamination in the suprapure salts and of course and the extreme values for the 29 lines measured in the to the presence of Mg itself as the matrix. Therefore, for Mg, diVerent Ca matrices. The values are given at 1300 W and at only the Mg II/I ratio is reported in Table 2.The Mg II/I the optimum rf power level for which the reference Cr II/I ratio suggests that the plasma is still robust for the Ca matrix. ratio is reproduced. At the optimum level, not only is the Its low value of 8.8 in the Mg matrix reflects another problem spread in the line intensity ratios at its minimum, but also the with the use of the Mg II/I ratio, especially when analysing matrix eVect has been removed. environmental samples. Owing to the high content and its The removal of the matrix eVect by power adjustments strong emission, the intensity of the Mg II line is out of the shows that the signal suppression mainly has its origin in the linear dynamic range.plasma conditions and that the sample introduction system hardly generates a suppressive eVect. The fact that all the The eVect of rf power adjustments solutions contained a similarly high acid level of 5% HNO3 may assist in the similar aerosol transport of the test samples As the reference, intensities were measured from a 10 mg l-1 or in hiding the eVect resulting from the sample introduction multielement standard in 5% HNO3 under the robust consystem.Even up to 1% of Ca, no matrix eVect is present, but ditions at an rf power level of 1300 W. Table 3 lists the 29 the spread in the line intensity ratios increases. Therefore, lines that were used. Then, the intensities were measured from although one may adjust the power level to compensate for the same multielement standard in 0.1, 0.3 and 1.0% Ca the Ca matrix eVect, matrix matching may be preferred when the salt content of the sample diVers widely from the salt Table 2 Chromium intensities (cps) and Cr and Mg ion-to-atom line content of the standards.ratios (not corrected for spectrometer eYciency) for diVerent salt In the set of 29 lines, the smallest eVects were observed for matrices of 1000 mg l-1 Cr I 357 nm, Al I 396 nm and Ga I 417 nm, whereas Zn II 202 nm, Zn II 206 nm and Cd II 214 nm showed the largest Line No salt Na K Mg Ca eVects.Obviously, the magnitude of the eVect corresponds Cr II 267 63200 62600 63800 60500 60200 with the atomic/ionic character of the line and its energy sum. Cr I 357 34500 34300 35000 34100 34600 The lines for Li, Na, Mg and K were used to determine the Cr II/I 1.83 1.83 1.82 1.77 1.74 line intensity ratios for the 0.1% Ca solution with respect to Mg II/I 12.0 12.0 12.0 8.8 11.7 the standard, but could not be used for the 0.3% and 1% Ca 834 J.Anal. At. Spectrom., 1999, 14, 831–838Fig. 4 Line intensity ratios with and without the presence of 1% Ca as a function of the energy sum of the 29 lines listed in Table 3. The dotted line indicates an intensity ratio equal to 1. Some lines discussed in the text have been specified. solutions owing to the presence of these elements as contami- Analysis of certified sediment samples nation in the suprapure calcium salt. In all of our measurements, the Mg II/I ratio was reproduced at a 10–50W higher In environmental analysis, samples such as soils and sediments are usually digested in 16% aqua regia and calibration stan- power as compared to the optimum power level based on the Cr II/I ratio (see Fig. 3 for an illustration). dards are accordingly matrix matched. Samples may vary in Fig. 5 Line intensity ratios for the 29 lines listed in Table 3. Intensities were measured from the 10 mg l-1 multielement standard in (a) 0.1% Ca and (b) 1.0% Ca in 5% HNO3.The reference intensities were measured from the same multielement standard in 5% HNO3 at 1300 W. Solid line: ratio for Cr II 267 nm; broken line: ratio for Cr I 357 nm. J. Anal. At. Spectrom., 1999, 14, 831–838 835Table 4 Averaged matrix eVect (%) and its extreme values (%) for the results in the same three values for the Zn content which equal 29 lines listed in Table 3, at 1300W and at the optimum rf power the reference value.The Cr results are also clearly improved. level (W) for which the reference Cr II/I ratio is reproduced in All line intensities increase on power adjustment, but the diVerent Ca matrices enhancement is stronger for the lines with a higher energy sum. When using the Mg II/I ratio instead of the Cr II/I ratio, At 1300 W At optimum power the rf power had to be increased even further. The Mg ratio Matrix Extreme Optimum Matrix Extreme is low at 1380W owing to the 200 mg l-1 Mg present in the eVect values power eVect values BCR277 solution.At this concentration level, the Mg II signal already suVers from non-linearity, whereas the reference signals 0.1% Ca -5 -10; 0 1350 -1 -3; 2 were determined from the 10 mg l-1 calibration standard. 0.3% Ca -7 -15; 0 1380 0 -3; 3 The eVect of rf power adjustments was also measured in the 1.0% Ca -17 -30; -5 1460 -2 -5; 4 axial viewing mode. The reference Cr II/I ratio in this mode diVers from the value in the radial mode, and was determined salt content between several hundreds to several thousands to be 0.835 in the multielement standard.It is not possible to of mg l-1 of Na, K,Mg and Ca. Typically, the ICP is supposed reproduce the Cr II/I ratio: in BCR277, values of 0.641 and to be robust with respect to this range in matrix composition 0.744 were measured at 1300 and 1500 W, respectively. On the and no severe matrix eVects are expected. However, based on contrary, the Zn 202/213 ratio was reproduced at 1500 W and the observations in the previous section, diVerent changes in the Zn 206/213 ratio at 1450 W.As the intensity of the Mg II intensity for various lines could be expected. line is beyond the range for BCR277, the Mg II/I ratio cannot In Table 5, the results are listed for the six elements of be determined. It is concluded that the rf power adjustments interest in the estuarine sediment BCR277, a reference material cannot minimize the spread in line intensity ratios as well as which is applied as a quality control sample in Dutch labora- in the radial viewing mode, although the robust conditions tories. External calibration matched for 16% aqua regia was can minimize the suppressions to the same extent as for radial used.For Cd and Pb, only one analyte line has been measured, viewing.7,22–24 In a very recent paper,28 however, observations as we do not intend to further complicate the present discussion similar to ours have been reported when comparing ionic lineby using prominent lines suVering from spectral interference based internal standardization in axial and radial viewing by Fe.In particular, the results for the three Zn lines measured modes to compensate for sodium eVects on accuracy. at 1300 W are remarkable: each line yields a diVerent content, By measuring the Cr II/I ratios for diVerent known salt and no result equals the indicative reference value. contents in radial mode and by adjusting the rf power level To investigate whether the salt content is responsible for for the reproduction of the reference ratio, the power adjustthese results, matrix matching was applied.Matrix matching ment as a function of the Cr II/I ratio can be determined to is possible for BCR277 because of the known salt content. control the matrix eVects. In a series of sediment samples with When about 1100 mg l-1 Ca and 200 mg l-1 Na, K and Mg unknown salt content (in particular, the amount of Ca and were added to the calibration standard, the results shown in Mg), all samples can be checked for their Cr II/I ratios.To Table 5 were obtained. The results for the three Zn lines are deviate ratios, new rf power levels are estimated from the the same and equal to the reference value. The results for Cr function and the corresponding samples must be rerun. This also improve. As outlined in the previous sections, robust idea is illustrated with four reference samples in Table 6: the conditions combined with acid matrix matching are not sewage sludge BCR145, the lake sediment BCR280 and the suYcient for this kind of analysis and the suppressions will be river sediments BCR320 and Maas.The intensities of Cr and due mainly to the Ca and Mg content. Zn at their prominent lines were measured. Calibration was In the aqua regia matched calibration standard and in the performed with the 10 mg l-1 multielement standard in 16% BCR277 solution, the Cr II/I ratio was determined.The two aqua regia at 1300 W. The samples which were digested in values are diVerent and the rf power requires an increase of 16% aqua regia were run under the same conditions. Based 80 W for the BCR277 solution to reproduce the reference on the measured Cr II/I ratios, the samples were rerun at the ratio. As can be seen from Table 5, this power adjustment optimum power. As can be seen from Table 6, the applied power levels are closely correlated to the Ca (plus Mg) Table 5 The concentrations found (mg g-1) for diVerent elements and concentrations specified for the reference materials. The results lines (nm) in the estuarine sediment BCR277.The indicative values for all samples clearly improve at all selected lines. In particular per element are included for the BCR145 and Maas samples, large power adjustments were made resulting in a much closer agreement for the Cr Concentration found 205, Cr 206 and Cr 267 lines with respect to the Cr 357 line as well as much less variation over the Zn lines.The results Power/W 1300 1300 1380 averaged over the lines are closer to the indicative values. External calibration at Intensities were also measured for the other lines listed in 1300 W matched for Acids Acids+salts Acids Table 5. The power adjustments induced a clear enhancement Element Energy and improvement for Ni and Pb in BCR145 and Maas, but line sum/eV Indicative only minor changes for the other sediments and elements.When analysing the series of sediment samples under robust Cd I 228 5.42 10.2 11.2 11.2 10.8 Cr II 205 12.8 130 140 142 145.6 conditions, several other actions may be taken.1 Matrix match- Cr II 206 12.8 125 132 136 ing, however, is not practical, as BCR280 and BCR320 require Cr II 267 12.9 130 138 140 a diVerent matrix composition of the standards than BCR145, Cr I 357 3.46 134 144 138 BCR277 and Maas. Standard additions are not attractive in Cu I 324 3.82 84 87 88 97.2 multielement analysis.In the present example, one has to spike Cu I 327 3.79 89 94 92 six diVerent elements in all the samples. When analysing Ni II 231 14.0 33 41 37 34.9 Ni I 232 5.34 35 46 37 unknown samples, an additional complication occurs, as, for Pb II 220 14.8 134 161 146 137.5 matrix matching, one has to know in advance the salt content Zn II 202 15.5 476 558 548 557 and, for standard additions, one should have at least an Zn II 206 15.4 489 553 562 idea about the analyte concentrations in order to add the Zn I 213 5.80 512 551 556 appropriate amounts of spikes. 836 J. Anal. At. Spectrom., 1999, 14, 831–838Table 6 The concentrations (mg g-1) found and indicative for several which the patterns of the lines are experimentally assessed for reference sediments. The external calibration was at 1300 W, matrix the actual ICP and spectrometer. When the library has been matched for 16% aqua regia. Samples were measured at 1300W and built, the samples must be measured under the same plasma at the optimum power level for which the Cr II/I ratio was reproduced.conditions. Salt matrix (mg l-1) is given for the measured solution The daily variations in the plasma parameters and the Concentration found sample matrix induce changes in the plasma conditions and, hence, in the line intensity ratios. As the strength of multiple At optimum Salt line procedures is said to be the independence of a prior At 1300W power Indicative matrix knowledge of the sample composition, matrix matching of standards is not feasible and large errors may occur if the BCR145 1420 W Na, 40 plasma conditions are not fully under control.39 As demon- Cr II 205 71 82 85.2±16.3 K, 80 Cr II 206 72 78 Mg, 350 strated in this work, the combination of robust ICP conditions Cr II 267 75 85 Ca, 2100 and rf power adjustments based on the Cr ion-to-atom line Cr I 357 79 85 ratio is an adequate alternative.If Cr happens to be absent in Zn II 202 2470 2760 2772±209 the unknown sample, other ion-to-atom ratios, such as those Zn II 206 2570 2870 for Mg or Pb, will perform almost as well or the sample Zn I 213 2650 2840 should be spiked with Cr. BCR280 1310 W Na, 350 Cr II 205 68 71 76±5 K, 540 Cr II 206 68 71 Mg, 320 Conclusions Cr II 267 70 73 Ca, 330 Cr I 357 71 73 The application of robust conditions is not suYcient to deal Zn II 202 263 275 290±16 with matrix eVects.Dissimilar line intensity changes are Zn II 206 275 286 observed in various salt matrices. Starting from the robust Zn I 213 278 290 BCR320 1330 W Na, 400 conditions, at which a reference value for the Cr II 267.716 nm Cr II 205 59 62 70.1±7.7 K, 490 to Cr I 357.869 nm line ratio is determined, rf power adjust- Cr II 206 60 62 Mg, 400 ments are made for samples with matrices up to 1% Ca to Cr II 267 58 62 Ca, 430 reproduce the Cr II/I ratio. At the optimum rf power level, Cr I 357 60 63 not only is the dissimilarity of the line intensities at a minimum, Zn II 202 107 123 124.4±5.4 but also the matrix eVect has been removed.For the 1% Ca Zn II 206 116 133 Zn I 213 119 131 matrix, no matrix eVect is observed from the sample Maas 1360 W Na, 10 introduction system. Cr II 205 153 164 169.2±13.2 K, 40 As compared to the Mg II/I ratio, the Cr II/I ratio varies Cr II 206 154 163 Mg, 190 more strongly with the coil distance and the rf power level, Cr II 267 153 164 Ca, 1080 making this ratio a better criterion for rf power adjustments.Cr I 357 160 160 The Mg II/I ratio is an appropriate indicator of the robustness Zn II 202 2520 2810 2740±135 Zn II 206 2590 2840 of the plasma, and a procedure to correct the measured ratio Zn I 213 2650 2820 for eYciency diVerences in the spectrometer is presented. The Ca matrix is used as the worst case example, as this matrix shows larger and more dissimilar suppressions for the 29 atomic and ionic emission lines studied as compared to Mg As the emission lines (also those of the same element) and, in particular, to Na and K matrices.experience diVerent suppressions, the use of one internal The practical application of power adjustments is illustrated standard, such as the Sc II 361.384 nm line23 or the Ar I with several certified sediment samples. The rf power adjust- 794.818 nm line,29 cannot compensate for the diVerent ments correspond to the Ca (and Mg) content of the sample responses of the lines to various concentrations of Ca.solutions. Results for several analytes and for diVerent lines Simulation of the change in energy transfer due to matrix of the same analyte improve. For instance, all three prominent mismatching reveals an almost continuous change in line Zn lines for BCR277 give identical results which equal the intensities as a function of the energy sum.8 A more or less reference value. The power adjustment approach does not linear change has indeed been observed by us, as illustrated in work properly in axial viewing, but is an attractive alternative Fig. 4. This indicates that it should be possible to describe this in radial viewing to the application of matrix matching, change by means of two or three lines covering the actual standard additions or internal standards. It is argued that energy sum range of the lines. It should even be possible to multiple line analysis in qualitative and semiquantitative use one atomic and one ionic line of the same element, e.g.analysis can benefit from power adjustments. In, Pd or Rh, on the Optima 3000 DV instrument. In view of the above discussion, rf power tuning based on the Cr II/I ratio is an attractive alternative to matrix matching, References standard additions or internal standardization. 1 D. A. Sadler, F. Sun, S. E. Howe and D. Littlejohn, Mikrochim. Acta, 1997, 126, 301. Multiple line analysis 2 A. Ferna�ndez, M.Murillo, N. Carrio�n and J.-M.Mermet, J. Anal. At. Spectrom., 1994, 9, 217. In qualitative and semiquantitative analysis, the determination 3 I. Novotny, J. C. Farinas, W. Jia-liang, E. Poussel and of whether an element is present is made from the observation J.-M.Mermet, Spectrochim. Acta, Part B, 1996, 51, 1517. of one or more of the most prominent lines of that element. 4 X. Romero, E. Poussel and J.-M. Mermet, Spectrochim. Acta, Part B, 1997, 52, 495. If the lines exist in the sample spectrum, the relative line 5 M.Carre�, K. Lebas, M. Marichy, M. Mermet, E. Poussel and intensities may be checked to confirm the element pres- J.-M.Mermet, Spectrochim. Acta, Part B, 1995, 50, 271. ence11,12,30,31 and to perform line selection.32,33 Or, mathe- 6 C. Dubuisson, E. Poussel, J. L. Todoli and J.-M. Mermet, matical procedures based on multiple linear regression,18,19 Spectrochim. Acta, Part B, 1998, 53, 593. principal component regression,34 target factor analysis35–37 7 J.-M.Mermet, J.Anal. At. Spectrom., 1998, 13, 419. and matrix projection38 can be applied. For multiple line 8 X. Romero, E. Poussel and J.-M. Mermet, Spectrochim. Acta, Part B, 1997, 52, 487. analyses, a tailor-made version of a library is necessary, in J. Anal. At. Spectrom., 1999, 14, 831–838 8379 J.-M.Mermet, Anal. Chim. Acta, 1991, 250, 85. 25 J. L. Todoli, J.-M. Mermet, A. Canals and V. Hernandis, J. Anal. At. Spectrom., 1998, 13, 55. 10 E. Poussel, J.-M. Mermet and O. Samuel, Spectrochim. Acta, Part 26 I. I. Stewart and J. W. Olesik, J. Anal. At. Spectrom., 1998, 13, B, 1993, 48, 743. 1249. 11 L. Soudier and J.-M. Mermet, Appl. Spectrosc., 1995, 49, 1478. 27 I. B. Brenner, M. Zischka, B. Maichin and G. Knapp, J. Anal. At. 12 C. Rivier and J.-M.Mermet, Appl. Spectrosc., 1996, 50, 959. Spectrom., 1998, 13, 1257. 13 L. M. Cabalin and J.-M.Mermet, Appl. Spectrosc., 1997, 51, 898. 28 C. Dubuisson, E. Poussel and J.-M. Mermet, J. Anal. At. 14 R. B. Bilhorn and M. B. Denton, Appl. Spectrosc., 1989, 43, 1. Spectrom., 1998, 13, 1265. 15 M. J. Pilon, M. B. Denton, R. G. Schleicher, P. M. Moran and 29 M. Hoenig, H. Doekalova� and H. Baeten, J. Anal. At. Spectrom., S. B. Smith Jr., Appl. Spectrosc., 1990, 44, 1613. 1998, 13, 195. 16 T. W. Barnard, M. I. Crockett, J. C. Ivaldi and P. L. Lundberg, 30 C. Trassy and J. Robin, Spectrochim. Acta, Part B, 1988, 43, 79. Anal. Chem. 1993, 65, 1225. 31 R. Neubo� ck, W. Wegscheider and M. Otto, Microchem. J., 1992, 17 T. W. Barnard, M. I. Crockett, J. C. Ivaldi, P. L. Lundberg, 45, 343. D. A. Yates, P. A. Levine, and D. J. Sauer, Anal. Chem. 1993, 32 D. P. Webb and E. D. Salin, J. Anal. At. Spectrom., 1989, 4, 793. 65, 1231. 33 D. P. Webb and E. D. Salin, Spectrochim. Acta, Part B, 1992, 18 E. H. van Veen, S. Bosch and M. T. C. de Loos-Vollebregt, 47, E1587. Spectrochim. Acta, Part B, 1997, 52, 321. 34 D. A. Sadler and D. Littlejohn, J. Anal. At. Spectrom., 1996, 19 J. A. Morales, E. H. van Veen and M. T. C. de Loos-Vollebregt, 11, 1105. Spectrochim. Acta, Part B, 1998, 53, 683. 35 D. F. Wirsz and M. W. Blades, Anal. Chem., 1986, 58, 51. 20 M. Thompson, M. H. Ramsey, B. J. Coles and C. M. Du, J. Anal. 36 D. F. Wirsz and M. W. Blades, J. Anal. At. Spectrom., 1988, At. Spectrom., 1987, 2, 185. 3, 363. 21 B. Budi and V. Hudnik, J. Anal. At. Spectrom., 1994, 9, 53. 37 D. F. Wirsz and M. W. Blades, Talanta, 1990, 37, 39. 22 C. Dubuisson, E. Poussel and J.-M. Mermet, J. Anal. At. 38 P. Zhang, D. Littlejohn and P. Neal, Spectrochim. Acta, Part B, Spectrom., 1997, 12, 281. 1993, 48, 1517. 23 I. B. Brenner, A. Zander, M. Cole and A. Wiseman, J. Anal. At. 39 E. H. van Veen and M. T. C. de Loos-Vollebregt, Spectrochim. Acta, Part B, 1998, 53, 639. Spectrom., 1997, 12, 897. 24 J. C. Ivaldi and J. F. Tyson, Spectrochim. Acta, Part B, 1995, 50, 1207. Paper 8/07979J 838 J. Anal. At. Spectrom., 1999, 14, 831

ISSN:0267-9477    

DOI:10.1039/a807979j

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Sample–standard interaction during trace analysis of semiconductorgrade trimethylindium by inductively coupled plasma atomic emission spectrometry Rajesh K. Gupta, Assad Al-Ammar and Ramon M. Barnes Department of Chemistry, Lederle Graduate Research Towers, University of Massachusetts, Box 34510, Amherst, MA 01003–4510, USA Received 8th January 1999, Accepted 10th March 1999 Trimethylindium (TMI), used to make III–V semiconductor compounds, was analyzed by inductively coupled plasma atomic emission spectroscopy.Several calibration approaches were evaluated for the determination of impurities. Oil-based standards dissolved in xylene and in 10% trimethylindium solution in xylene were tested for external and matrix-matched calibrations. Significant diVerences in the two calibrations were observed for most elements, causing severe determination error. This diVerence is due to an exchange between the indium in the sample matrix and the metal in the standard analyte compound (i.e., sulfonate, naphthenate, octanoate).An analyte species with enhanced volatility results. The analytes exhibiting highest enhancement are Al, Cd, Pb, Hg, Sn and Zn. Elements showing moderate enhancements are B, Ca, Mg, Mn and Si. Some analytes (i.e., Be, Fe) exhibit no signal enhancement. The sensitivity increased from several to several thousand per cent compared with standard analytes in xylene. Results from a flow merging experiment suggest that interaction between the matrix (TMI) and the standard analyte compound is fast and reaches completion in a few seconds. impurities by electrothermal vaporization (ETV)-ICP spec- Introduction trometry controlled with a suitable time–temperature program.Metal–organic chemical vapor deposition (MOCVD) is a pro- Decomposition methods have also been employed to convert cess that reacts vapor from Group II, III or IV metal alkyls with the volatile to non-volatile analyte compounds followed by ICPGroup V or VI hydrides to form semiconductor films.1 Gallium AES and ICP-MS analyses.However, some of the important arsenide (GaAs), for example, is grown from trimethylgallium analytes (e.g., Si) are hard to decompose and cannot be converted (TMG) and arsine (AsH3), AlGaAs from trimethylaluminium quantitatively to inorganic form.14 Also, the sample decompo- (TMA), TMG and AsH3 and InP from trimethylindium (TMI ) sition or hydrolysis is extremely exothermic and requires careful and phosphine (PH3).2 In this process, the high purity of the control even when performed at subambient temperatures.Thus, volatile, often pyrophoric, organometallic precursor material is volatile analyte compounds could be lost. important.3 Contaminating impurities in the grown semicon- Solvent dilution (SD) ICP-AES15 may be adapted for quality ductor material can aVect its conduction properties. Electronic control during TMI production. However, since the chemical properties of the layers can be destroyed if undesirable impurity forms of the elemental impurities are unknown, calibration is levels exceed 1015 free carriers cm-3, corresponding to approxi- not entirely reliable if standards are prepared with non-volatile mately one part in 107.4 Furthermore, as semiconductor device organic acid salts.16 Hence, when applied to TMI analysis, matrix-matched calibration should be useful to correct for spec- fabrication techniques continue toward high component densitroscopic and non-spectroscopic matrix interferences.We sus- ties, low-level contamination becomes increasingly important pected, nonetheless, that the highly reactive TMI might interact and its determination challenging.5,6 with standard analyte compounds to convert them to chemical Using contamination-free starting materials can insure forms that exhibit diVerent volatility behavior from the starting reliable and durable semiconductor devices.7,8 Trimethylcompound.As a result, an error in determination can occur indium is the preferred precursor in indium-based alloy fabriowing to the diVerence in chemical structure (and hence vola- cation by MOCVD.9 In semiconductor wafer fabrication, a tility) between standard and impurity analyte compounds. high-purity, inert gas (e.g., hydrogen, helium) delivers TMI to In this study, chemical interactions were investigated a deposition chamber where it reacts with PH3. This carrier between TMI and the organometallic compounds usually gas is saturated with TMI vapor as it bubbles through a TMI employed for commercial standard preparation.A qualitative sample container. Determining the volatility of contaminants estimation of the determination error is proposed. For this and quantifying them is especially important since in the purpose, diVerent calibration approaches were evaluated for MOCVD process mostly volatile contaminants are carried to SD-ICP-AES.Numerous analyte lines were studied and the the reaction chamber and non-volatile components remain in eVect of the TMI matrix is discussed. the reagent bubblers.10 Several approaches have been described to distinguish volatile from non-volatile impurities and to quantify each. Bertenyi and Experimental Barnes11 developed an exponential dilution technique that Instrumentation involves the analysis of a totally vaporized sample. Volatile impurities at the chamber temperature were determined by ICP- A simultaneous multi-element ICP-AES system (Optima 3000, Perkin-Elmer, Norwalk, CT, USA) was used for the analysis AES.Argentine et al.12,13 speciated volatile and non-volatile J. Anal. At. Spectrom., 1999, 14, 839–844 839Table 2 Analytical lines listed by wavelength of trimethylindium samples. The instrumental parameters are described in Table 1 and the analytical wavelengths studied Analytical line and Excitation+ionization Soft (S) or are listed in Table 2.wavelength/nm Linea potentials/eV hard (H) line Sampling Zn 202.551 II 6.1 S Cr 205.552 II 6.0 H The trimethylindium samples (Lot 810008, Morton Si 212.415 I 6.6 S International, Metalorganics Division, North Andover, MA, Zn 213.856 I 5.8 S USA) are reactive and pyrophoric, and special care is required Cd 214.438 II 14.7 H Pb 220.350 II 14.7 H while handling them. A glove-box (Blickman, Weekhawken, Cu 224.700 II 15.9 H NJ, USA) purged with dried, de-oxygenated nitrogen was Cd 226.502 II 14.4 H used to prepare and dilute samples in an inert atmosphere.Cd 228.802 I 5.4 S The arrangement for sample handling and storage is illustrated Be 234.861 I 5.4 S in Fig. 1. A Teflon-coated spatula was used to transfer a Sn 235.484 I 6.2 S sample aliquot from the shipping containers into 120 mL glass Fe 238.204 II 13.0 H Fe 239.562 II 13.0 H bottles (Qorpak, 4 oz flint, Cat. No. 36319–765; VWR Sn 242.949 I 5.9 S Scientific, Westchester, PA, USA).An appropriate amount B 249.773 I 4.9 S was diluted in xylene to prepare test samples and matrix- Si 251.612 I 4.9 S matched standards. Hg 253.652 I 4.9 S Mn 257.610 II 12.2 H Reagents Fe 259.940 II 9.8 H Mn 260.569 II 12.2 H Standard solutions for 16 elements (Ag, Al, B, Be, Ca, Cd, Pb 261.417 I 5.7 S Cu, Fe, Mg, Mn, Pb, Sn, Sr, Zn, Hg and Cr) were prepared Be 265.062 I 7.4 S Cr 267.716 II 6.2 H by dissolving hydrocarbon oil-based standards (VHG Mg 279.553 II 12.0 H Laboratories, Manchester, NH; at 100 mg g-1) in xylene.All Mg 280.270 II 12.0 H standards except Cu, Mn, Sn, Hg and Cr were made from Pb 283.307 I 4.4 S dialkyl arylsulfonates in hydrocarbon oil. Copper and Cr were Sn 283.999 I 4.8 S naphthenate compounds, Mn was an octanoate compound Mg 285.213 I 4.3 S and Hg was a sulfonate compound. Tin was a tetravalent Si 288.159 I 5.1 S Al 309.283 I 4.0 S organotin compound. The oil-based standards were dissolved Be 313.042 II 13.2 H in (i) m-xylene (anhydrous grade, 99% purity; Aldrich, Cu 324.754 I 3.8 S Milwaukee, WI, USA) and (ii) xylene containing 10% m/m Cu 327.396 I 3.8 S pure trimethylindium (Morton International ).The former Ag 328.068 I 3.8 S served as the standards for external calibration and the latter Zn 334.502 I 7.8 S were used for matrix-matched calibration. Ag 338.289 I 3.6 S Cd 361.050 I 7.3 S Ca 393.366 II 9.2 H Method Al 396.152 I 3.1 S Ca 396.847 II 9.2 H Solvent dilution. Two sets of standard solutions were pre- Mn 403.076 I 3.1 S pared for calibrations.One set was prepared to study cali- Sr 407.771 II 8.7 H bration by dissolving and preparing the oil-based standards Sr 421.552 II 8.6 H in xylene solvent. Pure m-xylene was used as a blank. aI are atom lines and II are ion lines. Organometallic elemental standards at 0.25 and 0.5 ppm were prepared by suitable dilution of a 100 ppm primary standard. These contained all 16 elements. Table 1 Operating conditions for ICP-AES analysis ICP system Optima 3000 prototype Rf power 1.3 kW Frequency, free running 40 MHz ICP torch Type 2 quartz slotted extension Torch injector Ceramic alumina Outer argon flow rate 15 L min-1 Intermediate argon flow 2.0 L min-1 rate Central argon flow rate 0.8 L min-1 Observation zone 15 mm above coil Fig. 1 Glove-box arrangement with purged, dried and deoxygenated Observation height ±2.5 mm nitrogen used for all sample preparation steps. Nebulizer Concentric glass (Glass Expansion, Hawthorn, Australia) Sample pump rate 0.8 mL min-1 The second set, which represents the matrix-matched stan- Pump tubing Viton, orange–orange dard solutions, was prepared by dissolving 16 organometallic (id 0.035 in) element standards in m-xylene that contained 10% m/m TMI. Spray chamber Glass Scott double-pass, coolant The solvent with 10% m/m TMI matrix was prepared first and jacketed (Spectro Analytical used for diluting the standards to make up 0.25 and 0.5 ppm Instruments, Fitchburg, MA) Spray chamber temperatures 0 °C (flow merge); standard solutions.Pure xylene with 10% m/m TMI was used 10 °C (calibration) as a blank. The standard for Si, prepared from bis(trimethyl- Integration time (auto) 10 s (maximum) silyl )methane (Gelest, Tullytown, PA, USA), was separately Background correction ±0.04 nm added to make up 0.25 and 0.5 ppm m/m, since it had to be Drain Pumped prepared fresh. Other elements remain stable in solution over 840 J.Anal. At. Spectrom., 1999, 14, 839–844an extended time (~1–2 years). Calibration functions were evaluated for individual elements prepared in these two matrices. The spray chamber temperature was maintained at 10 °C when calibration plots were determined. Merging flow solvent dilution. To characterize the interaction of TMI and analyte compounds, a merging flow experiment was designed (Fig. 2). Two sample introduction channels were merged just before the nebulizer–spray chamber. Four experiments were performed to observe the signal response.In experiment O, only xylene flowed through both channels. The corresponding signal served as the blank for all analytes. In experiment A, a standard in xylene flowed in channel 1 and pure xylene flowed through channel 2. The resulting signal corresponded to standards prepared in only xylene. In experiment C, standard solutions prepared in 10% TMI matrix flowed through channel 1 and pure xylene flowed through channel 2.The signal response corresponded to standards prepared in TMI matrix. In experiment B, standards in xylene flowed through channel 1 and 10% TMI matrix solution flowed through channel 2. The signal corresponded to the standard made on-line in 10% TMI matrix solution. Little or no matrix– standard interaction was assumed to occur, since little time had passed before the sample reached the discharge. In all cases, the dilution factor was two since the peristaltic pumps 1 (channels 1 and 2) and 2 (nebulizer feed) were maintained 0.25 and 0.5 mL min-1, respectively.Owing to the highly reactive, pyrophoric nature of TMI, the entire arrangement was housed in a glove-bag (Model X-27–27, Instruments for Research and Industry, Cheltenham, PA, USA) purged with N2 during this entire experiment. Results and discussion Calibrations Calibrations with xylene and matrix-matched standards were compared to establish whether they could be used for the accurate analysis of semiconductor grade TMI. If the slopes of the two calibration functions were the same, then dissolving and analyzing organometallic standards in xylene would be suYcient for the analysis of TMI samples.The results summarized in Table 3, however, show significant diVerences between the two calibration series. The slope of each spectral line reflects the corresponding sensitivity. Except for Ag and Cu, the sensitivity for all elements was greater for the matrix-matched 10% TMI in xylene standards than for the xylene standards.A large sensitivity increase is observed for Al, Cd, Hg, Sn, Pb and Zn between calibration with the standards dissolved Fig. 2 Experimental set-up for the merging flow experiment with two peristaltic pumps in a nitrogen purged glove-bag. Peristaltic pump 2 operates at half the pump rate of pump 1, and excess flow is diverted to waste. Table 3 Calibrations by standards in xylene and in 10% TMI matrix Calibration by standards in xylene Calibration by standards in 10% TMI Standard deviation DLa in Standard deviation DLa in Element spectral Slope/counts Intercept Correlation of blank xylene solid TMI Slope/counts Intercept Correlation of 10% TMI in xylene solid TMI line/nm (ppm)-1 (counts) coeYcient (r2) (counts) (ppm) (ppm)-1 (counts) coeYcient (r2) (counts) (ppm) Ag I 328 2450 103 0.9999 1.9 0.02 242 55.9 0.969 21 2.6 Al I 396 513 58.5 1.000 11 0.6 1596 530 0.9999 17.7 0.3 B I 249 448 -21.3 0.9996 4.3 0.3 795 5.3 0.9998 2 0.08 Be II 313 16954 1819 0.9969 5.5 0.01 17614 388 1.000 10.8 0.02 Ca II 393 169794 13996 0.9979 24 0.004 264752 6468 0.9998 29 0.003 Cd I 228 16.4 3.6 1.000 0.1 0.15 63 1.7 0.9936 2.7 1.3 Cr II 267 114.9 19.8 0.9976 3 0.8 194 786 0.9982 27 4.2 Cu I 327 12001 21.7 1.000 8.5 0.02 867 387 0.9519 7 0.24 Fe II 259 219 -11.3 0.9996 2.1 0.3 233 -7.3 0.9996 2.8 0.4 Hg I 253 19.6 50 0.9789 0.8 1.2 146 48.9 0.9974 3.5 0.7 Mg II 279 4047 434 0.9944 5.7 0.02 8302 -124 0.9999 0.4 0.001 Mn II 257 308 14.6 0.9992 1.7 0.2 472 -19.1 0.9997 0.3 0.02 Pb I 261 28.6 -2.8 0.9677 1.1 1.2 289 -18.8 0.9997 7.3 0.7 Si I 251 96.6 8.1 1.000 1.4 0.4 165 34.4 0.9954 0.9 0.2 Sn I 283 25.3 7.2 0.9998 1.1 1.3 181 -55.6 0.9885 0.9 0.15 Sr II 407 73460 9539 0.9986 62 0.02 112337 8207 0.9996 138 0.004 Zn I 213 93.2 15.4 1.000 0.6 0.2 852 288 0.9973 19 0.7 aDetection limit (DL) calculated as three times the standard deviation of the blank/calibration slope and multiplied by 10 since 10% solutions of TMI samples were analysed.J. Anal. At. Spectrom., 1999, 14, 839–844 841in xylene and those with 10% TMI in xylene (Table 3). The If the TMI matrix reacts with the standard compound to form more volatile compounds (or a mixture of compounds), the Pb 261 nm and Zn 213 nm lines show ten- and nine-fold increases in sensitivity, respectively, in the TMI matrix com- analyte nebulization eYciency should increase considerably.Consequently, a higher signal is expected because more analyte pared with the xylene matrix. The two Zn line calibrations are shown in Fig. 3, and the two Zn function intercept values reaches the plasma. To establish the presence of and study the nature of a matrix–standard interaction, matrix and standards diVer by approximately 20-fold. This indicates the presence of a Zn impurity in the TMI matrix. Ultra-pure trimethylindium were mixed just before nebulization in a merging flow arrangement (Fig. 2).The spray chamber temperature was maintained was not available to prepare matrix-matched standards. The Hg 253 nm and Sn 283 nm lines were enriched more than at 0 °C instead of 10 °C when calibration plots were determined. The lower temperature should decrease the volatile compound seven-fold and the Cd 228 nm and Al 396 nm lines showed a 3–4-fold sensitivity increase. Similar intercept values indicate signals as the TMI matrix reacts with the standard analyte compounds.Therefore, the signals from the standards in the no Cd and Hg impurities were present in the TMI matrix, although the Al function intercept suggests a major impurity. TMI matrix and in xylene should not diVer as much as during the calibration experiment (Table 3). The results from the Modest sensitivity increases of 1–2-fold were observed for B, Si, Mn, Cr, Mg, Ca and Sr. The Si 251 nm line showed a merging flow experiment verify this expectation (Fig. 4).The histograms in Fig. 4 represent normalized signals if the ca. 70% increase in sensitivity in the TMI matrix compared with the calibration in xylene. The small diVerence in intercepts signals from the xylene standards are unity. Experiment A represents the signals from standards in xylene. Experiment B corresponds to 1–2 mg Si g-1, depending on the calibration used to calculate the concentration. This level of Si is usually the concentration limit in TMI and similar metal alkyls used for semiconductor fabrication.No significant diVerence in sensitivity was found for the Fe 259 nm and Be 313 nm lines. Copper and Ag were the only elements that showed a decrease in sensitivity in the presence of the TMI matrix. Coupled with the intercept and standard deviation of the blank, the slope is inversely proportional to the limit of detection, which is also listed in Table 3. These sensitivity diVerences can be attributed to (i) a change in plasma conditions upon TMI matrix introduction, (ii) an interaction between the standard compound and the TMI matrix resulting in the formation of a more volatile species or (iii) a combination of (i) and (ii).A number of experiments were undertaken to test these possibilities. Matrix-standard interactions Aqueous solutions. The hypothesis tested first was a change in plasma conditions with indium matched standards. An aqueous solution of 10% m/m ultra-pure In metal containing 5 mg mL-1 of all the analytes was introduced into the plasma to study its eVects on analytical line sensitivity.When compared with the analyte signals for the same concentration without an In matrix, signal suppression and enhancement were observed ranging between +50% and -50% depending on whether the analytical lines were atomic or ionic. This change is not suYcient to explain the large diVerences in calibrations observed, some of which are as much as 2000%.Flow experiments. The hypothesis that volatile species are formed in TMI-matched standard solutions also was examined. Fig. 4 Normalized signals for (a) Ag, Al, B, Be, Ca and Cd, (b) Cr, Cu, Fe, Mg and Mn and (c) Hg, Pb, Sn, Sr and Zn for merging flow experiments. Experiment A represents the signals from standards in xylene. Experiment B represents the signal from TMI matrix standards that are mixed and prepared on-line. Experiment C represents the Fig. 3 Zn 213 nm emission line calibration curves.&, Calibration in signal from the standards prepared in TMI matrix before the experiment. TMI matrix; 2, calibration in xylene. 842 J. Anal. At. Spectrom., 1999, 14, 839–844represents the signal from standards in the TMI matrix pre- dards can be attributed to some well defined properties of pared on-line. Experiment C represents the signal from the Group III metal alkyls. standards prepared in the TMI matrix before the experiment. Any reaction between a standard analyte compound and Transmetalation.Group III metal alkyls such as TMA, TMI should be complete in experiment C, because suYcient TMG and TMI sometimes undergo a metal or metal ion time had elapsed before the standards were measured. exchange when they react with metal halides, carboxylic acids However, in experiment B the TMI mixes with the standard salts and sulfonic acids salts: analyte compound prepared in xylene after pump 1, and MR+Me3In�InR+Me3M therefore the standards in the TMI matrix are prepared on-line.The mixed sample reaches the spray chamber about 5 s after The removal of metal M from its salt of an organic acid, MR, merging. Therefore, the reaction between the TMI and the results in a non-volatile In compound and a volatile species, standard organometallic analyte compounds is expected to be Me3M. The reaction mechanism and the nature of this either incomplete or absent. The signal, when compared with exchange are well understood.17,18 In the present study the that from experiment C, should indicate the reaction rate.metal ion from the standard compound is assumed to replace In Fig. 4, the Al, B, Ca and Cd signals are enhanced when In in TMI to form a volatile species. Alkyl reactions and experiment B and C results are compared with experiment A. rearrangements also could result in mixed metal alkyls and Lead, Sn and Zn are strongly enhanced in the presence of the the formation of monomethyl, dimethyl or tetramethyl metal TMI matrix.The interaction rate with TMI is fast for Pb and compounds depending on the oxidation number of the metal. Zn and comparatively slow for Sn. However, in time Sn forms volatile species, resulting in a large signal increase in the TMI Mixed alkyl complexes.19,20 The Group III metal alkyls matrix. Aluminium is one of the major elements showing strong prefer chemical forms that have mixed alkyl groups. For signal enhancement; the Al 309 and 396 nm lines are enhanced example, 1-methyl-1-ethyl-1-propylindium is a more stable almost as much in experiment B as in experiment C.Thus chemical form than trimethylindium. interaction with the TMI matrix is almost complete (~90%) R1R2RM+Me3In�Me2RM+R1R2MeIn in a few seconds. Calcium and Cd signals are enhanced in the TMI matrix, as experiment C shows. Their TMI–standard The alkyl (or aryl ) groups, R1 and R2, in an organic salt, interaction rate is fast and nearly complete in a few seconds.R1R2RM, with metal M are replaced by the methyl groups This interaction is slower with boron, but eventually a volatile in TMI, resulting in a volatile species Me2RM (or R2MeRM, species results. R1MeRM). Since most of the standard compounds used in The silver (328 and 338 nm lines) signals are enhanced our experiment are dialkyl arylsulfonates, some of the alkyl because of volatile compound formation. However, after a or aryl groups are exchanged between the standard and the few hours, silver starts to precipitate in the solution, resulting trimethylindium resulting in the volatile standard compound.in lower signals from experiment C. Copper, like Ag, precipi- Metal alkyl compounds of Al, B, Cd, Hg, Pb, Si, Sn and tates from the solution. The small signal enhancement in Zn are volatile and can be expected as reaction products. No experiment B is due to the matrix loading, as observed earlier metal alkyls form with Ag, Be, Ca, Cu, Fe, Mn or Sr that when the eVect of a 10% aqueous In solution on analytical might account for the insignificant sensitivity and signal lines was studied. Iron exhibits a small signal suppression changes observed.typical of plasma matrix loading. For Be no enhancement was These compounds have not been identified experimentally, observed in either experiment B or C. Some signal suppression although gas chromatography atomic emission detection can be attributed to plasma matrix loading.This corroborates (GC-AED)21–23 measurements are planned. Tests with organo- the results from the calibration experiment (Table 3) wherein metallic antimony, arsenic, bismuth, germanium, selenium, the Be sensitivity did not change in the TMI matrix. Chromium tellurium and thallium standards are expected also to produce also shows no aYnity to form a volatile species rapidly volatile reaction products and enhanced sensitivity in a TMI with TMI. matrix-matched standard.The Mg and Mn analyte lines show no enhancement (experiment C). However, when the calibration experiment was performed at a higher spray chamber temperature (10 °C), Conclusions Mg and Mn showed an enhancement ranging from 60 to Experimental results indicate that some of the analyte salts 100%. The possible reason could be the formation of a species (Al, Cd, Hg, Pb, Sn and Zn) used to prepare elemental (or mixture of species) that is volatile compared with their standards in organic solvents strongly interact with trimethyl- standard form, dialkyl arylsulfonates, but not suYciently indium and result in more volatile analyte compounds.Other volatile to appear at a low spray chamber temperature (0 °C). elements (B, Ca, Mg, Mn and Si) show similar but less A novel feature of the Mg and Mn lines studied is that the Mg 285 nm line, the only soft line (279 and 280 nm are hard ounced trends. As a result of this interaction, elements lines), and the Mn 403 nm line, the only hard line (257 and such as Ag and Cu are not very stable in solution and 260 nm are soft lines), show opposite trends to their other eventually precipitate. lines in experiment B.However, the diVerence is small A suitable calibration is required to achieve reliable and (~15–20%). This may be due to the observation height accurate analysis of TMI and similar metal alkyls. A calioptimization for soft lines, since they constitute the majority bration that uses standards chemically similar to the impurity of all the analytical lines studied. The hard lines are not as analyte compound is more suitable than one that uses nonsensitive in this observation zone, and therefore they do not volatile organometallic salts as standards.Although the chemireflect exactly the same trend as soft lines. cal forms of the impurity are unknown, they are presumed to These results demonstrate that volatile species are generated be volatile, because the semiconductor grade TMI is purified from organometallic standards in the presence of TMI.Since by successive sublimation. Hence the matrix-matched calimany metal alkyls are volatile, their formation upon reaction bration is expected to give results that are more reliable. with TMI could account for the signal enhancements observed. However, a suitable method that eliminates volatility eVects would be ideal. Volatility corrections required for ICP-AES Reaction mechanisms and ICP mass spectrometry with volatile organic samples have been demonstrated by Al-Ammar et al.24 to improve accuracy These experimental interactions have a theoretical basis.The reaction between trimethylindium and organometallic stan- for TMI and other materials analyses. J. Anal. At. Spectrom., 1999, 14, 839–844 8436 L. Rothman, D. Quinlan and S. Koch, in Proceedings of the 1992 The volatility of the standard compound after reaction with Microcontamination Conference, Santa Clara, CA, pp. 635–644. the TMI matrix can be exploited to enhance the detection 7 World Wide Web site http://rel.semi.harris.com/docs/lexicon.html limits. The spray chamber temperature should be optimized, 8 World Wide Web site http://rel.semi.harris.com/docs/lexicon/ so that most desolvation takes place without losing much of manufacture.html the volatile standard analyte compound. Improved detection 9 R. J. Malik, Semiconductor Materials and Devices, Elsevier, Amsterdam, 1989.limits for B, Si, Mn, Mg, Sn and Sr were obtained in this way. 10 R. IscoV, Semicond. Int., 1990, July, 70. For other elements not much improvement resulted, because 11 I. Bertenyi and R. M. Barnes, Anal. Chem., 1988, 58, 1734. the impurity in the TMI used to make the matrix-matched 12 M. D. Argentine, A. Krushevska and R. M. Barnes, J. Anal. At. standards resulted in high blank values. Spectrom., 1994, 9, 1121. 13 M. D. Argentine and R. M. Barnes, J.Anal. At. Spectrom., 1994, 9, 1371. Acknowledgements 14 K. Takeda, M. Minobe, T. Hoshika, T., Jinno and T. Yako, Analyst, 1990, 115, 535. The authors thank Perkin-Elmer (Norwalk, CT, USA) for 15 R. I. Botto and J. J. Zhu, J. Anal. At. Spectrom., 1994, 9, 905. providing the Optima-3000 prototype system and Morton 16 A. C. Lazar and P. B. Farnsworth, Anal. Chem., 1997, 69, 3921. Metalorganics (North Andover, MA, USA) for supplying 17 D. S. Matteson, Organometallic Reaction Mechanisms of the Non- samples and reagent chemicals. This investigation was sup- Transition Elements, Academic Press, New York, 1st edn., 1974, ported by ICP Information Newsletter, Inc. (Hadley, MA, p. 36. USA). 18 A. W. Parkins and R. C. Poller, An Introduction to Organometallic Chemistry, Macmillan, London, 1st edn., 1986. 19 A. J. Pearson, Metallo-Organic Chemistry, Wiley, New York, 1st References edn., 1985, ch. 3. 20 P. Powell, Principles of Organometallic Chemistry, Chapman and 1 A. Ohsawa, K. Honda, R. Takizawa, T. Nakanishi, M. Aoki and Hall, London, 2nd edn., 1988. N. Toyokura, Proc.-Electrochem. Soc., 1990, 90-7 (Semicond. 21 S. A. Estes, P. C. Uden and R. M. Barnes, Anal. Chem., 1981, Silicon 1990), pp. 601–613. 53, 1829. 2 T. Shimano, M. Morita, Y. Muramatu and M. Tsuji, in 22 K. J. Slatkavitz, L. D. Hoey, P. C. Uden and R. M. Barnes, Anal. Proceedings of the Sixth Workshop in ULSI and Ultra Clean Chem. 1986, 57, 1846. Technology, December 20, 1990, pp. 59–68. 23 P. C. Uden, J. Chromatogr. A, 1995, 703, 393. 3 S. A. Koch and T. Pinkston, ICP Inf. Newsl., 1994, 20(7), 491. 24 A. S. Al-Ammar, R. K. Gupta and R. M. Barnes, J. Anal. At. 4 A. C. Jones, P. R. Jacobs, R. CaVerty, M. D. Scott, A. H. Moore Spectrom., in the press. and P. J.Wright, J. Cryst. Growth, 1986, 77 (1–3), 47. 5 P. Van Zant, Microchip Fabrication: a Practical Guide to Semiconductor Processing, McGraw Hill, New York, 1990. Paper 9/00275H 844 J. Anal. At. Spectrom., 1999, 14, 839–844

ISSN:0267-9477    

DOI:10.1039/a900275h

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

Speciation of arsenic in fish tissue using microwave-assisted extraction followed by HPLC-ICP-MS† Kathryn L. Ackley, Clayton B’Hymer, Karen L. Sutton and Joseph A. Caruso* Department of Chemistry, University of Cincinnati, P.O. Box 0172, Cincinnati, OH 45221–0172, USA Received 24th September 1998, Accepted 22nd February 1999 The use of microwave-assisted extraction for the extraction of arsenic species from fish tissue is described. Quantitative extraction of arsenic from spiny dogfish muscle (CRM, DORM-2) was achieved using methanol–water (80+20, v/v) with microwave heating at 65 °C in a closed-vessel microwave system. Extractions were performed with a variety of solvents including water, two diVerent methanol–water mixtures, and a 5% tetramethylammonium hydroxide solution. Extracted arsenic species were separated using both ion-exchange and ion-pair chromatography with ICP-MS detection.The DORM-2 along with three diVerent varieties of fish purchased from a local market were analyzed for arsenic. In all samples, the majority of arsenic present was in the form of arsenobetaine, a nontoxic arsenic species.arsenic species of interest in seafood samples. Inductively Introduction coupled plasma mass spectrometry (ICP-MS) has been used The determination of total arsenic is not suYcient to assess extensively as a detector for arsenic speciation.12,15–19,21 the risks associated with consumption of arsenic-containing ICP-MS oVers high sensitivity but is hindered in the detection foodstuVs since the toxicity of arsenic is highly dependent on of arsenic by a polyatomic interference caused by ArCl+.its chemical form with inorganic arsenic (arsenite and arsenate) However, this interference may be minimized by using being more toxic than monomethylarsonic acid (MMA) and HPLC.16,24 Chromatography may separate the chloride ions dimethylarsinic acid (DMA). Arsenobetaine and arsenocho- from the analytes of interest.Therefore, the chloride ions line are relatively non-toxic. As a result, much attention has would elute in a single peak instead of being present continubeen given to the elemental speciation of arsenic in environ- ously in the background and leading to ArCl+ in the plasma. mental and biological samples, and the subject has been No matter how eVective the separation technique is or how reviewed by Burguera and Burguera.1 sensitive the detector, the quality of the analysis is limited by Arsenobetaine has been identified to be the major arsenical the sample preparation step.Extraction of the analytes from in a variety of seafood products including several species of the tissue sample is usually achieved using a methanol– clams2 and many species of fish.3–6 Arsenobetaine is thought water–chloroform system.5,25,26 Although these solvent extracto be the final metabolite of arsenic in marine food chains,7 tion systems have been shown to extract arsenic from fish although arsenobetaine is not present in all species of fish,8 tissue quantitatively, the procedure is time consuming.In and the transformations that arsenic undergoes in the marine addition, extraction of the analytes from certain types of food chain are still being studied.7–10 samples such as mussels25 may be less eYcient. To perform arsenic speciation analyses, extraction methods Extraction of arsenic using a conventional solvent extraction must be capable of quantitatively extracting arsenic from the method is time consuming.Microwave-assisted extraction sample while not altering the individual arsenic species in any (MAE) is another alternative to conventional solvent extraction. way. Typically, arsenic species are extracted from seafood In MAE, microwave energy is used to heat solvents that are in samples using a standard solvent extraction method. Methanol, contact with solid samples so that analytes of interest will water and sometimes chloroform are used as the extracting partition from the sample into the solvent.27 Open focused solvents, and the procedure is performed at ambient tempera- microwave systems have been utilized for the dissolution, tures and pressures.Branch et al.3 detailed a typical method extraction and derivatization of organotin compounds in biomafor the extraction of arsenic from fish tissue. The sample and terials28–30 and sediments30,31 prior to speciation. Open focused extracting solvents were sonicated for 1 h.The solution was microwave-assisted sample preparation procedures have also centrifuged, the supernatant collected and the process repeated. been investigated for the application to speciation of mercury The methanol–water layer was separated from the chloroform in environmental samples.30 All these papers report that analytes layer. The solvent was removed by rotary evaporation, and were eVectively extracted, while the species information the sample dissolved in water.remained intact. Also, the overall analysis time was reduced The approaches to arsenic speciation in seafood have been dramatically with the use of the microwave systems. varied. Separation of the analytes is usually achieved through Microwave digestion has been used to prepare seafood32 high-performance liquid chromatography (HPLC). Ion-pair and certified marine samples33,34 for the determination of total chromatography11–15 as well as ion-exchange chromatogra- arsenic.However, in both cases, the digestion conditions were phy15–23 have successfully been employed for separating the harsh and no attempt was made to retain the integrity of each individual species. Two reports of the use of MAE for the preparation of †Presented at the Ninth Biennial National Atomic Spectroscopy Symposium (BNASS), Bath, UK, July 8–10, 1998. samples for arsenic speciation were found. Demesmay and J. Anal. At. Spectrom., 1999, 14, 845–850 845Olle�35 utilized microwave digestion to prepare sediment obtained from the United States Environmental Protection Agency (US EPA), Cincinnati, OH, USA.The compounds samples for the speciation of arsenic. The sediment samples were exposed to microwave power in the presence of a were used without further purification. MMA (ChemServices, West Chester, PA, USA), DMA (as cacodylic acid, Sigma, St. hydrochloric and nitric acid mixture. Of the four species investigated, MMA, DMA and As(V) were all stable during Louis, MO, USA), sodium arsenite and sodium arsenate (Matheson, Coleman and Bell, Norwood, OH, USA) were the microwave digestion.Arsenite [As(III)] was quantitatively oxidized to As(V). Larsen et al.36 utilized microwave heating also used without further purification. MMA and DMA were in the form of free acids. for the extraction of arsenic species from dried mushroom samples. Methanol–water (1+9, v/v) was used as the Samples extracting solvent.The purpose of the work reported here was to develop a The certified reference material was DORM-2 (National microwave extraction procedure capable of quantitatively Research Council of Canada), which has a certified arsenic extracting arsenic from seafood samples using extraction conconcentration of 18.0±1.1 mg kg-1. The market fish samples ditions that are mild enough to prevent a conversion of the were purchased as fresh filets from a local grocery store.original arsenic species. A variety of extracting solvents were Where required, the outer scales and skin of the fish were investigated. The eVect of varying the exposure time and removed with plastic utensils. The samples were then lyophil- temperature was also investigated. ized and homogenized. All samples were analyzed as freeze-dried powders. Experimental Extraction procedure Equipment Samples of DORM-2, ranging in mass from 0.10 to 0.13 g, The ICP-MS instrument was a VG PQ2 STE (VG Elemental, were accurately weighed into the reaction vessels.The Franklin, MA, USA) equipped with a concentric nebulizer extracting solvent (10 mL) was added to each sample. The and a double pass spray chamber, water-cooled to 5 °C. The microwave system was programmed to heat the sample to a rf power was 1350 W, and the nebulizer gas flow was optimized specified temperare, and the temperature was then main- prior to analysis with a 10 mg L-1 In solution. tained until 2 min had elapsed.Thus, if a sample was to be An MES 1000 (CEM, Matthews, NC, USA) microwave heated for 6 min, the 2 min sequence was initiated three times. extraction system was used. This is a closed-vessel system After microwave heating, the samples were allowed to cool capable of heating 12 samples at one time. The temperature and were then transferred into centrifuge tubes. The samples was monitored in a control vessel by an armored fiber-optic were centrifuged for 5 min at 3500 rpm.The supernatants were temperature control probe, and the pressure was monitored transferred into sample vials using Pasteur pipets (Fisher by an internal pressure control system. Samples were heated Scientific). Before injection, the samples were filtered with a to a set temperature by the microwave system, and microwave 0.2 mm Nylon disposable syringe filter (Fisher Scientific). energy was applied to the sample when its temperature fell below the set temperature.The reaction vessels were Advanced Flow injection analysis (FIA) Composite Vessels supplied by CEM. The liner and liner covers of these vessels were Teflon PFAA and have a maximum For flow injection work, the mobile phase consisted of 2% v/v operating temperature and pressure of 200 °C and 200 psi, nitric acid solution prepared with ACS plus nitric acid (Fisher respectively. Vessel liners were cleaned first using soap and Scientific) and filtered prior to use. The mobile phase was water followed by soaking for 1 h in a 50% nitric acid solution.pumped at a rate of 1 mL min-1. All standards were prepared The liners were rinsed with distilled, de-ionized water, then in the same extracting solvent as the samples to be analyzed. 2% nitric acid, and finally, more distilled, de-ionized water. Data were collected using time-resolved software, so both m/z An ISCO Model 2350 HPLC pump and an ISCO Model 2360 75 (75As and 40Ar35Cl ) and m/z 77 (77Se, 40Ar37Cl ) could be gradient programmer (ISCO, Lincoln, NE, USA) were used monitored.Both masses were monitored to determine the to deliver the mobile phase for both the flow injection and extent of ArCl+ formation. All calibration graphs and subchromatographic work. Samples were injected for flow injec- sequent sample concentrations were determined using peak tion and chromatography using a Rheodyne Model 4396 areas. The arsenic concentrations were determined from three injector (Rheodyne, Cotati, CA, USA) equipped with a 20 mL replicate samples, each analyzed in triplicate. sample loop.A 47 cm length of PEEK tubing (0.020 id, 0.062 od) was used to connect the injector to the nebulizer Chromatography during flow injection experiments. The same tubing was used Anion-exchange separations were performed using a Hamilton to connect the chromatographic column to the nebulizer during PRP X100 column (250×4.1 mm id) with a 10 mm particle HPLC-ICP-MS experiments. size (Phenomenex, Torrance, CA, USA).The mobile phase consisted of a 30 mmol ammonium carbonate (J.T. Baker, Reagents and standards Phillipsburg, NJ, USA) buVer. The pH of the solution was The water used was distilled and de-ionized to a resistance of adjusted with ammonia solution (Fisher Scientific) until a pH 18 MV cm. Arsenic standards for the total arsenic determi- of 9 was reached. The mobile phase flow rate was 1 mL min-1. nations were prepared using a 1000 mg mL-1 As stock solution The ion-pair separation was performed with a C18 column (Spex CertiPrep, Metuchen, NJ, USA).The diVerent extrac- (15 cm×2.0 mm id) with a 5 mm particle size (Phenomenex). tion solvents were prepared with distilled, de-ionized water, The mobile phase was an aqueous 25 mM citric acid solution HPLC-grade methanol (Fisher Scientific, Fair Lawn, NJ, (Fisher Scientific) that was also 10 mM in 1-pentanesulfonic USA), and 25% tetramethylammonium hydroxide solution acid sodium salt (Eastman Kodak, Rochester, NY, USA). (Alfa Aesar, Ward Hill, MA, USA).Solutions of 1000 mg L-1 The mobile phase flow rate was 0.15 mL min-1. (as As) of each of the arsenic species investigated were prepared. These standards were stored at ambient temperature in Discussion the dark. The arsenobetaine and arsenocholine were synthesized by Dr. William R. Cullen (University of British Detection of arsenic with quadrupole-based ICP-MS can be problematic if chloride ions are present in the sample since Columbia, Vancouver, British Columbia, Canada) and were 846 J.Anal. At. Spectrom., 1999, 14, 845–850Table 1 Comparison of extraction solvents for MAE. Samples were At first, these results seem counter intuitive, and one would heated to 50 °C for 4 min. Arsenic concentrations are reported expect more arsenic to be extracted at the higher temperature in mg kg-1. DORM-2 has a certified As value of 18.0±1.1 mg kg-1 of 80 °C.One possible explanation for this phenomenon may be derived from the boiling-point of the extraction solvent. At Determined As 50 °C, the extraction solvent may not be suYciently heated to Extraction solvent concentration/mg kg-1a RSD (%) produce convection currents that agitate the tissue sample and Distilled, de-ionized water 13.4±0.4 2.9 aid in the liberation of the arsenic species. Most conventional 5% Tetramethylammonium 17.1±0.9 6.6 solvent extraction techniques involve sonication or some form hydroxide solution of mechanical agitation.The samples prepared using MAE do Methanol–water (50+50) 15.5±0.8 6.4 not receive the same type of agitation. Samples were placed in Methanol–water (80+20) 13.4±1.2b 10.6 a carousel that slowly revolved to promote even heating. aConfidence limits were determined at a 95% probability interval. However, the speed of rotation was not high and would not be bAverage value based on eight measurements. The ninth value was expected to agitate the sample within the microwave vessel.At rejected by Q-test with >95% confidence. 65 °C, the sample is close enough to the boiling-point so that suYcient agitation occurs to extract the arsenic species from the tissue quantitatively. At the higher temperature of 80 °C, ArCl+ has the same mass-to-charge ratio as As+ (m/z 75). To the solution boils vigorously when observed in an open vessel. ascertain if the presence of ArCl+ would interfere with the Hence, one would expect a greater portion of the extracting detection of arsenic, mass 77 was also monitored since approxisolvent to reside in the gas phase at this elevated temperature, mately one quarter of all the ArCl+ would be expected to leaving less liquid to eVect the extraction.This may explain have a mass-to-charge ratio of 77. No appreciable signal was why the amount of arsenic extracted is lower at 80 °C than obtained at m/z 77. Therefore, all the counts obtained at m/z at 65 °C. 75 were attributed to arsenic ions. Fig. 1 shows the results from an experiment to compare the A comparison of extraction solvents was performed, and amount of arsenic extracted from samples exposed to micro- these results are reported in Table 1. All samples were heated waves and from those not exposed to microwaves. The samples to 50 °C and exposed to microwave power for 4 min. At this were prepared in the same manner as described previously. In temperature and exposure time, the highest recovery was both cases, samples were placed in the microwave carousel for obtained with an extraction solvent of 5% tetramethylamthe same period of time.However, only one set of samples monium hydroxide (TMAH). A TMAH solution was initially was exposed to microwave power. While a large portion of investigated because other investigators have used TMAH to the arsenic is recovered from the tissue using methanol–water solubilize tissue in an open focused microwave system.29 The (80+20, v/v) without microwave exposure, microwave heating high recovery is not unexpected since the TMAH solution ensured that all the arsenic was extracted from the sample.digests and solubilizes the tissue. Only a very small amount of Slightly more arsenic is extracted without microwave heating tissue residue was visible after the sample had been centrifuged. using the methanol–water solution than with microwave heat- This extraction solvent was not investigated further since other ing at 50 °C, illustrating that no real advantage is seen with neutralized TMAH solutions were found to shift peak retention MAE until the sample is heated to a temperature at the times when injected onto the C18 column.In addition, the boiling-point of the extracting solvent. Similarly, no real retention times in subsequent analyses were also shifted even diVerence was seen between microwave-heated and unheated if the sample was prepared in a diVerent solvent.samples when water was used as the extracting solvent at 65 °C. A univariate approach was used to optimize the temperature Pure water extracted the same amount of arsenic as methand exposure times. The first extraction solvent investigated anol–water (80+20, v/v) with greater precision at 50 °C. Since was methanol–water (80+20, v/v) since methanol–water solu- the analysis of arsenicals in distilled, de-ionized water would tions are commonly used to extract arsenicals from tissue at be ideal for introduction into chromatographic systems, an ambient temperature and pressure.Table 2 shows the arsenic investigation of the amount of arsenic extracted by water at recovered from DORM-2. various temperatures and exposure times was performed. The The largest variation in the amount of arsenic extracted is data are presented in Table 3. observed when samples are heated to 65 °C. About 80% of the The trends that were prominent in the data obtained using arsenic is extracted from the samples heated to 50 °C.The methanol–water as the extraction solvent were not visible in samples heated to 80 °C show slightly better recoveries. the data reported in Table 3. Average values increased slightly However, samples heated to 65 °C eVectively extract 100% of with increasing exposure time at 50 and 65 °C but decreased the arsenic in DORM-2. The pressure inside the reaction vessel slightly at 80 °C. Moreover, the average amount of arsenic fluctuates slightly at 65 °C but typically is below 10 psi.At extracted increased slightly with increasing temperature for 80 °C, the pressure remains above 10 psi and may fluctuate as high as 20 psi when microwave power is being applied. Table 2 Arsenic recovered from DORM-2 using MAE with methanol –water (80+20, v/v) as the extraction solvent. Arsenic concentrations are reported in mg kg-1. DORM-2 has a certified As value of 18.0±1.1 mg kg-1 Determined As concentration (mg kg-1)a at: Exposure time 50 °C 65° C 80°C 2 min 13.9±0.6 20.5±0.5 17.6±0.8 4 min 13.4±1.2b 19.4±0.5 14.2±0.8 6 min 11.9±0.4 18.7±0.4 13.8±0.6 aConfidence limits were determined at a 95% probability interval.bAverage value based on eight measurements. The ninth value was Fig. 1 Comparison between the amount of arsenic extracted from rejected by Q-test with >95% confidence. DORM-2 with and without microwave heating. J. Anal. At. Spectrom., 1999, 14, 845–850 847Table 3 Arsenic recovered from DORM-2 using MAE with distilled, de-ionized water as the extraction solvent.Arsenic concentrations are reported in mg kg-1. DORM-2 has a certified As value of 18.0±1.1 mg kg-1 Determined As concentration (mg kg-1)a at: Exposure time 50 °C 65° C 80° C 2 min 12.7±0.3 13.6±0.3 15.1±0.2 4 min 13.4±0.4 13.8±0.2 14.5±0.7 6 min 13.8±0.2 14.7±0.2 14.4±0.3 aConfidence limits were determined at a 95% probability interval. Fig. 2 Chromatogram of arsenic species separated on the Hamilton PRP-X100 anion-exchange column. Each species is present at a concentration of 17 ng mL-1 As, and the sample was prepared in most samples.This trend was more prominent for samples methanol–water (80+20, v/v). Peak identification: (1) arsenocholine, (2) arsenobetaine, (3) arsenite, (4) dimethylarsinic acid, (5) monome- heated for only 2 min. It is possible that 2 min is not a thylarsonic acid, (6) arsenate. suYcient length of time for the arsenic species to be totally extracted from the fish tissue. As a result, samples heated for only 2 min show more dramatic increases in the amount of arsenic extracted with increasing temperature.heated to 100 °C as well as the increased variation between the replicate samples. Results from other workers3,15,34 (ref. 3 and 15 refer to DORM-1 while ref. 34 refers to DORM-2) as well as those After total arsenic concentration information had been obtained with FIA-ICP-MS, speciation information was gath- presented here (working with DORM-2) show that the arsenic in DORMexists almost exclusively in the form of arsenobetaine.ered using HPLC-ICP-MS. Two diVerent chromatographic separations were used to identify the arsenic species. Anion- Hence, the MAE method with water, which extracts 70–80% of the arsenic in the certified reference material, is capable of exchange chromatography was utilized for the identification of the most toxic forms of arsenic (arsenite, arsenate, MMA, extracting arsenobetaine but not extracting it quantitatively.This phenomenon suggests that an equilibrium process of some and DMA). Arsenobetaine and arsenocholine co-elute in the void volume. A sample chromatogram is shown in Fig. 2. The sort may be taking place, in which case the use of a larger volume of extracting solvent may be advisable if the concen- use of a gradient between buVers of low and high ionic strength was investigated for the separation of these species. tration of arsenic in the sample is large enough so that excessive dilution will not hinder detection.The other possible explanation While using gradient elution provides better resolution between the first three peaks, gradient elution alters the plasma charac- is that the arsenobetaine may be incorporated within the sample matrix in more than one way. However, no reliable studies of teristics and necessitates re-equilibration between runs.Hence an isocratic separation, which provides suYcient resolution arsenobetaine binding in fish have been performed. The amount of arsenic recovered with water under the for the identification of each anionic species, was employed. A mobile phase with a pH of 9 was selected since, at this pH, conditions investigated is not as large as that obtained with a mixture of methanol and water. One possible explanation is the anionic species of interest have diVerent apparent charges and are more easily separated on the anion-exchange column that, since water has a higher boiling-point, the temperatures investigated did not produce adequate convection currents used in this work.37 To identify the species eluting in the void volume, samples were also separated using ion-pair chromatog- within the sample to provide suYcient agitation to aid in extraction.raphy. While reversed-phase techniques such as ion-pair chromatography tend to be more sensitive to matrix eVects than An investigation was made to see if more arsenicals could be extracted with water at higher temperatures.Samples of anion-exchange separations, preparing standards in the same extracting solvent as the samples to be analyzed allows the DORM-2 were prepared using water as the extracting solvent. The samples were heated for 1 min to 65 °C and then for 4 min identification of organoarsenicals by their retention times. Any sample preparation method to be used in conjunction to 100 °C.The amount of arsenic extracted in the three replicate samples was determined to be 16.4±1.8 mg kg-1. with speciation must not alter the individual species within the sample. To determine if this criterion was met, standard Although the average arsenic concentration for the three replicates is lower than the certified value, they cannot be solutions of each species were prepared in methanol–water (80+20, v/v). The test standards were placed in the microwave distinguished at a 95% confidence level. This result seems to support the idea that extractions performed near the boiling- system and exposed to microwave power for 4 min at 65 °C.The control standards were not exposed to microwave power, point of the extracting solvent extract arsenicals from the fish tissue more completely. and the solutions were analyzed using the anion-exchange chromatographic method with ICP-MS detection. The chrom- The variation in the amount of arsenic extracted between the three samples prepared in water at 100 °C is larger than atograms indicate that no sample degradation occurred, and the integrity of the species remained intact. While this result the variation obtained for the other samples prepared with water at lower temperatures.The appearance of these samples was expected for the organoarsenic species, the fate of the inorganic forms of arsenic was unknown since Demesmay and is also diVerent from that of samples heated at lower temperatures.Samples prepared at lower temperatures essentially Olle�35 reported the conversion of arsenite to arsenate in certain extraction/digestion media. However, the extraction conditions appeared the same as they did prior to microwave heating. Samples heated to 100 °C appeared much darker and had a in our work were not severe enough for oxidation to occur. Once the method had been shown to maintain species integ- much stronger odor. In addition, large portions of the sample were found to be adhered to the upper inside walls of the rity, the fish tissue extracts were analyzed using HPLC-ICP-MS.Anion-exchange chromatography was the first separation tech- TeflonA microwave vessel liners. The temperature may be high enough at 100 °C to go beyond heating the sample and may nique utilized, and the resulting chromatograms are shown in Fig. 3(a)–(d). All the fish extract chromatograms exhibited very actually begin to ‘cook’ the sample.This phenomenon would account for the diVerent odor and appearance of the samples large peaks where arsenobetaine and arsenocholine co-elute. 848 J. Anal. At. Spectrom., 1999, 14, 845–850Fig. 4 (a) Ion-pair chromatogram of arsenic species in water. Peak identification: (1–3) arsenite, arsenate, and MMA, (4) DMA, (5) arsenobetaine, (6) arsenocholine. (b) Ion-pair chromatogram of arsenic species in 80% methanol. Peaks elute in the same order as in (a).Fig. 4(a). The retention times are shifted when the species are introduced in methanol–water (80+20, v/v) as shown in Fig. 4(b). However, since arsenobetaine was expected to be the primary arsenic species in each fish sample, the separation was adequate for species confirmation. The ion-pair chromatograms for the fish samples are shown in Fig. 5(a)–(d). These chromatograms confirm that arsenobetaine accounts for most of the arsenic in DORM-2 while it accounts for all of the arsenic in the ocean whitefish, black tip shark and steelhead salmon samples.A small peak matching the retention time of DMA was observed in the anion-exchange chromatogram for DORM-2. The chromatogram is consistent with work previously reported in the literature. Goessler and co-workers34 (working with DORM-2) reported an arsenobetaine concentration of 16.0±0.7 mg kg-1, a DMA concentration of 0.28±0.01 mg kg-1, and inorganic arsenic, MMA and arsenocholine concentrations were all reported to be lower than 0.03 mg kg-1. Peaks for inorganic arsenic, MMA and arsenocholine were not observed in this work, which may be a result of diVerences in sensitivity between this method and other methods reported in the literature.An attempt was made to quantify the arsenobetaine using Fig. 3 (a) Anion-exchange chromatogram of DORM-2 extract in 80% HPLC-ICP-MS. The results are presented in Table 4. However, methanol. (b) Anion-exchange chromatogram of ocean whitefish these values should not be taken as an absolute quantification extract in 80% methanol.(c) Anion-exchange chromatogram of black since the arsenobetaine standard solutions used to generate tip shark extract in 80% methanol. (d) Anion-exchange chromatogram of steelhead salmon extract in 80% methanol. the calibration graph were prepared from arsenobetaine of unknown purity. Less than 1 g of arsenobetaine was available for this analysis, so further purification and standardization was not feasible. Solutions or arsenobetaine and arsenocholine Inorganic forms of arsenic were not detected while the DORM-2 chromatogram indicated a small peak where DMA elutes.were analyzed by HPLC-ICP-MS, and only one peak was observed for each standard. The presence of other substances, Any uncharged or positively charged organoarsenical would be expected to be unretained on the anion-exchange column, such as waters of hydration, was not known, making the preparation of a standard solution for quantification diYcult.so to confirm the presence of arsenobetaine, an alternative chromatographic technique was utilized. When the arsenic As the demand for accurate arsenic speciation information in food and environmental samples grows, so does the need for species are injected onto the C18 column as an aqueous solution, adequate resolution is obtained between DMA, commercially available standards of the major arsenic species of interest. arsenobetaine, and arsenocholine.The separation is shown in J. Anal. At. Spectrom., 1999, 14, 845–850 849Acknowledgements The authors gratefully acknowledge the US EPA for providing arsenic samples, lyophilizing fish samples and for funding through grant number CX826144–01–0. We also thank the CEM Corporation for the use of the MES 1000 unit. Although this work has been funded with EPA funds, it has not been subjected to the Agency’s peer and policy review and does not reflect the views of the EPA.References 1 M. Burguera and J. L. Burguera, Talanta, 1997, 44, 1581. 2 W. R. Cullen and M. Dodd, Appl. Organomet. Chem., 1989, 3, 79. 3 S. Branch, L. Ebdon and P. O’Neill, J. Anal. At. Spectrom., 1994, 9, 33. 4 E. H. Larsen, G. Pritzl and S. H. Hansen, J. Anal. At. Spectrom., 1993, 8, 1075. 5 D. Beauchemin, M. E. Bednas, S. S. Berman, J. W. McLaren, K. W. M. Siu and R. E. Sturgeon, Anal. Chem., 1988, 60, 2209. 6 E. H. Larsen, G. A. Pedersen and J.W. McLaren, J. Anal. At. Spectrom., 1997, 12, 963. 7 W. Goessler, W. Maher, K. J. Irgolic, D. Kuehnelt, C. Schlagenhaufen and T. Kaise, Fresenius’ J. Anal. Chem., 1997, 359, 434. 8 J. S. Edmonds, Y. Shibata, K. A. Francesconi, R. J. Rippingale and M. Morita, Appl. Organomet. Chem., 1997, 11, 281. 9 K. Hanaoka, T. Motoya, S. Tagawa and T. Kaise, Appl. Organomet. Chem., 1991, 5, 427. 10 W. R. Cullen, L. G. Harrison, H. Li and G. Hewitt, Appl. Organomet. Chem., 1994, 8, 313. 11 X.C. Le, M.Ma and N. A. Wong, Anal. Chem., 1996, 68, 4501. 12 X. C. Le and M. Ma, J. Chromatogr. A, 1997, 764, 55. 13 P. Boucher, M. Accominotti and J. J. Vallon, J. Chromatogr. Sci., 1996, 34, 226. 14 X.-C. Le, W. R. Cullen and K. J. Reimer, Talanta, 1994, 41, 495. 15 D. Beauchemin, K. W. M. Siu, J. W. McLaren and S. S. Berman, J. Anal. At. Spectrom., 1989, 4, 285. 16 C. Demesmay, M. Olle and M. Porthault, Fresenius’ J. Anal. Chem., 1994, 348, 205. 17 S. Saverwyns, X. Zhang, F.Vanhaecke, R. Cornelis, L. Moens and R. Dams, J. Anal. At. Spectrom., 1997, 12, 1047. 18 E. H. Larsen, Fresenius’ J. Anal. Chem., 1995, 352, 582. 19 P. Terasahde, M. Pantsar-Kallio and P. K. G. Manninen, J. Chromatogr. A, 1996, 750, 83. Fig. 5 (a) Ion-pair chromatogram of DORM-2 extract in 80% 20 K. J. Lamble and S. J. Hill, Anal. Chim. Acta, 1996, 334, 261. methanol. (b) Ion-pair chromatogram of ocean whitefish extract in 21 S. Branch, K. C. C. Bancroft, L. Ebdon and P. O’Neill, Anal. 80% methanol. (c) Ion-pair chromatogram of black tip shark extract Proc., 1989, 26, 73. in 80% methanol. (d) Ion-pair chromatogram of steelhead salmon 22 Z. Mester and P. Fodor, J. Anal. At. Spectrom., 1997, 12, 363. extract in 80% methanol. 23 J. Mattusch and R.Wennrich, Anal. Chem., 1998, 70, 3649. 24 B. S. Sheppard, W.-L. Shen, J. A. Caruso, D. T. Heitkemper and F. L. Fricke, J. Anal. At. Spectrom., 1990, 5, 431. 25 M. B. Amran, F. Lagarde and M. J. F. Leroy,ikrochim.Acta, Table 4 Quantification of arsenic species in fish samples. Arsenic 1997, 127, 195. concentrations are reported in mg kg-1.a Extractions were performed 26 M. Ochsenkuhn-Petropulu, J. Varsamis and G. Parissakis, Anal. at 65 °C for 4 min using methanol–water (80+20, v/v)b Chim. Acta, 1997, 337, 323. 27 H. M. Kingston and S. J. Haswell, Microwave-Enhanced Sample Arsenobetaine Total arsenic determined by FIA Chemistry, American Chemical Society, Washington, DC, 1997, p. 772. DORM-2 24.6±1.4 19.4±0.5 28 I. R. Pereiro, V. O. Schmitt, J. Szpunar, O. F. X. Donard and Steelhead salmon 2.9±2.0 2.6±0.1 R. Lobinski, Anal. Chem., 1996, 68, 4135. Black tip shark 19.9±1.9 14.5±0.5 29 V. O. Schmitt, J. Szpunar, O. F. X. Donard and R. Lobinski, Can. Ocean whitefish 7.9±4.5 7.2±0.1 J. Anal. Sci. Spectrosc., 1997, 42, 41. aConfidence limits were determined at a 95% probability interval. 30 V. O. Schmitt, A. d. Diego, A. Cosnier, C. M. Tseng, J. Moreau bFor chromatographic determinations, n=2–3. and O. F. X. Donard, Spectroscopy, 1996/1997, 13, 99. 31 O. F. X. Donard, B. Lale�re, F. Martin and R. Lobinski, Anal. Chem., 1995, 67, 4250. 32 B. S. Sheppard, D. T. Heitkemper and C. M. Gaston, Analyst, 1994, 119, 1683. 33 G. Damkro� ger, M. Grote and E. Janssen, Fresenius’ J. Anal. Conclusions Chem., 1997, 357, 817. Microwave-assisted extraction is an eVective method for the 34 W. Goessler, D. Kuehnelt, C. Schlagenhaufen, Z. Slejkovec, and K. J. Irgolic, J. Anal. At. Spectrom., 1998, 13, 183. extraction of arsenic species from fish tissue samples. When 35 C. Demesmay and M. Olle�, Fresenius’ J. Anal. Chem., 1997, 357, methanol–water (80+20, v/v) was used as the extracting 1116. solvent, each of the species investigated remained intact when 36 E. H. Larsen, M. Hansen and W. Go� ssler, Appl. Organomet. exposed to microwave power for 4 min at 65 °C. Under these Chem., 1998, 12, 285. conditions, 100% of the arsenic in DORM-2 was extracted. In 37 J. Gailer and K. J. Irgolic, Appl. Organomet. Chem., 1994, 8, 129. all of the fish samples investigated, the arsenic existed almost exclusively in the form of arsenobetaine. Paper 8/07466F 850 J. Anal. At. Spectrom., 1999, 14, 845&n

ISSN:0267-9477    

DOI:10.1039/a807466f

出版商:RSC    

年代:1999

数据来源: RSC