|
1. |
Information retrieval from the inductively coupled plasma |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 99-108
Barry L. Sharp,
Preview
|
|
摘要:
Information retrieval from the inductively coupled plasma† Plenary Lecture Barry L. Sharp,*a Jonathan Batey,b Ian S. Begley,‡a David Gregson,§b John Skilling,c Azli B. Sulaiman¶a and Griet Verbogta aDepartment of Chemistry, Loughborough University, Loughborough, UK LE11 3TU bVG Elemental, Ion Path, Road Three, Winsford, Cheshire, UK CW73BX cMaxEnt Solutions Ltd., Whipsiderry, 14 Lowfields, Little Eversden, Cambridge, UK CB3 7HJ Received 24th September 1998, Accepted 30th November 1998 This paper describes various means by which information can be derived from the signals obtained from inductively coupled plasmas.The first part describes how Bayesian analysis may be used to infer the species contributing to observed spectra and to estimate the contributions of those species to the intensity in each spectral channel. An example is given in which a synthetic mixture of rare earths (Nd, Sm, Eu and Gd), in the presence of a high level of Ba, is analysed and an inferred spectrum synthesized that fits the observed data to within three standard deviations at all spectral positions. The second part deals with the use of correlated measurements as a means of reducing noise and improving the precision of analytical measurements.The use of rapid peak jumping in quadrupole MS to emulate simultaneous measurement is shown to oVer significant improvements in the determination of lead isotope ratios with precisions of the order of 0.04% being achieved.Optical detection of ICP spectra has been revolutionized by the introduction of array detectors. It has been reported that simultaneously measured line and near background signals have a degree of mutual correlation and this results in noise reduction on their subtraction. Under ideal conditions the removal of correlated noise yields improvements in detection limits of about one order of magnitude. The final part describes how the ICP and other atomic spectrometric sources can be employed to obtain molecular information.The two basic approaches are to employ prior separation of the molecular species or to use tunable ionization sources whose conditions can be varied so as to provide both atomic and molecular information. A specific example of prior separation given here is the coupling of capillary electrophoresis to ICP-MS and its use for the speciation of CrIII and CrVI and Ni–humic acid complexes. analysis, is it becoming possible to extract this information Introduction from ICP signals.The field of analytical atomic spectrometry is in a state of Modern analytical science is about providing information change and the rate of that change appears to be increasing. that is fit for its intended purpose. Analytical measurements The driving force is undoubtedly the developing requirement do not provide such information directly, rather they provide to determine both elemental concentrations and molecular data.These data are not a complete or faithful representation form in a wide variety of sample types. Already, and increas- of the real world, but an approximation that may be biased ingly, the distinction between atomic and molecular analysis and is certainly modified by the presence of noise. Given the is becoming blurred. A goal of analytical science must be to data, we infer from them what the real world is like and then combine the low limits of detection, accuracy and precision of through a model information is derived.atomic spectrometric techniques with a capability for providing The ICP is, so far, unchallenged as a provider of information structural and chemical information such as that provided by on elemental content, requiring only a linear calibration model other forms of chemical spectroscopy. How does the ICP to link the data and concentrations. Looking more closely, shape up in this scheme of things? The ICP is a wonderful however, there are problems.There is noise associated with example of the reductionist approach to problem solving in the data that limits the precision. When the data from ICP that it reduces virtually everything that is thrown at it to the experiments are plotted on wavelength or mass scales, it is lowest common chemical denominators, the elements. Even found that more than one species can contribute to the counts so, the ICP provides a huge amount of information about a in any particular measurement channel.This becomes critical sample and only now, through the use of multivariate data when the number of species exceeds the number of independent channels available for their estimation, that is, the problem is under-constrained. Further, the sensitivity for diVerent species may be matrix dependent. These factors combine to scramble †Presented at the Ninth Biennial National Atomic Spectroscopy Symposium (BNASS), Bath, UK, July 8–10, 1998. the data obtained from the ICP, particularly for complex ‡Present address: Scottish Crop Research Institute, Invergowrie, samples, and means must be found to recover the desired Dundee, UK DD4 6RA.information. §Present address: Applied Photophysics, 203–205 Kingston Road, Molecular information is destroyed by the ICP when it is Leatherhead, Surrey, UK KT22 7PB. operated under standard conditions. However, using it as an ¶Present address: Department of Chemistry, Faculty of Science, element specific detector coupled to separation techniques, its Universiti Teknologi Malaysia, KB 791, 80990 Johor Bahru, Johor, Malaysia.reductionist qualities are again exploited to provide definitive J. Anal. At. Spectrom., 1999, 14, 99–108 99information on elemental content. Alternatively, the ICP or for another element will have the essential property of orthogonality that permits successful regression. Ideally, other discharges can be used under ‘soft’ conditions where molecules that do not fragment too easily can be kept intact.the higher order PCs should be zero, but inevitably they are not. Noise will cause small departures from zero and This paper describes recent work in our laboratory and elsewhere that has attempted to extract the maximum deviations significantly above the noise level indicate the presence of interferences. Sadler and Littlejohn6 described information from the ICP and other atomic spectrometric sources. Specific examples include spectral deconvolution a method which uses the higher order PCs either to reject interfered with lines or to estimate the contribution from employing probabilistic (Bayesian) methods; noise reduction through the use of correlated measurements, e.g., for isotope interference in the measured data.Ultimately, such methods are limited when spectral interferences are pre- ratio measurements; and coupled techniques such as capillary electrophoresis-ICP-MS. The use of soft ionization sources sent on all the available wavelengths or when multiplicative interferences destroy the orthogonality of the that conserve molecular information will also be discussed.components. (iv) Negative concentrations may be reported because zeros 1 Spectral decomposition employing Bayesian in the spectrum can be interpreted as linear combinations methods of positive and negative contributions. (v) The presence of noise (s) has an indeterminate eVect on Modern analytical chemistry is almost universally concerned the outcome.with estimating chemical or physical quantities from instru- (vi) No reliable estimate of the error is provided. mental outputs. However, these outputs are at best only partial Nevertheless, such methods have been used, but depend on representations of the input. The input is usually an additive expert knowledge to condition the matrix to provide good combination of positive components that are distributed with data.A key point is that this knowledge is applied reluctantly respect to some common variable. Such additive sets are and ad hoc to avoid disaster. formally referred to as measures.1 The most common example Bayesian methods1 start from a diVerent perspective and encountered in analytical chemistry is a spectrum, which may treat the true spectrum M, the measure, as a probabilistic be distributed in wavelength, frequency, energy or mass. rather than an absolute quantity. Similarly for the data, D.Conventionally, the measure is treated as an absolute quan- Provided that the instrument response function [the point tity that is recovered from the data through calibration of the spread function (PSF)] and the noise are characterized, the instrument response with known standards. Thus, the intensity output produced by a particular measure can be readily at one point in the spectrum may be represented by calculated. This output is referred to as the likelihood and is i=Srjcj+b+s (1) expressed as a probability: prob(D given M). For example, an elemental mass spectrum can be synthesized by multiplying 92 where rj is the system response factor (sensitivity) for the jth elemental concentrations by the individual isotopic distribution component, cj is the concentration of the jth component, b is patterns (which here constitute the PSF), and the instrument the background intensity and s is the noise.Generally and responses, to give a spectrum of 238 data points (this ignores necessarily (if more than one component is contributing to i ), the molecular ions, but the isotopic patterns for these can also many spectral elements are observed, thus a matrix is generbe included if desired).Assuming a Poisson noise distribution, ated. Hence the probability of obtaining a data point of magnitude d, I=RC+B+s (2) when the true mean value of the measure is m, is given by where C is the vector representing the concentrations of the prob (d given m)=e-m md/d! (3) individual components.If the number of spectral elements measured is equal to the number of unknowns, there is only Thus, multiplying the probabilities for all the data points gives prob(D given M). The previously known information about one solution, R-1 (I-B), which is obtained by inverting the matrix. the spectrum is encapsulated in the PSF, which has diVerent forms depending on the application. Here it is the isotopic However, several problems can arise: (i) If the data are under-constrained, R-1 does not exist.patterns, but in an NMR spectrum it could be a function representing the individual lineshape, e.g., a Lorentzian profile. (ii) If the data are over-constrained, there are many possible solutions and a least squares approach is usually employed The data that are generated by convoluting the unknown terms (e.g., the concentrations) with the PSF (and the cali- to find the solution that minimizes the sum of the squares of the diVerences between the actual and predicted data.brated responses) are known as the mock data. However, when an unknown spectrum is recorded, to esti- It should be noted, however, that the solution for which the number of data points and components is equal nearly mate the measure inferred from the data is a non-trivial problem. This output quantity is referred to as the posterior: always (depending on the magnitude of the noise) gives the best fit, even if some of the components are not prob(M given D).The two probabilities are linked by the product rule. The joint probability is given by actually present in the sample. (iii) An assumption in regression analysis is that there is no correlation between the observed variables. That is, there prob(M and D)=prob(M)×prob(D given M)=prob(D)×prob(M given D) F F F F is no additive or multiplicative interference. In the case where correlations do exist, multivariate methods such as prior likelihood predictive posterior (4) principal components regression (PCR) and partial least squares (PLS) can be employed.2–7 These attempt to find linear combinations of the variables (the PCs) that can The predictive prob(D) is a number that indicates how eVectively the calculated measure fits the data.It is arrived at by account for all the variability related to changes in one of the unknowns. For example, in an elemental optical integrating the posterior over the whole space of possible measures, i.e., spectrum containing several un-interfered with spectral lines of one element, a principal component formed from the combined measurements at these wavelengths will prob(D)=Sprob(D and M)=Sprob(M)×prob(D given M) all possible M all possible M contain all the variability due to concentration variations in that element. A principal component formed similarly (5) 100 J.Anal. At. Spectrom., 1999, 14, 99–108In practice, it is not possible to explore all the possible this is that the preferred solution is the one that maximizes disorder.The entropy10 of the system is defined as solutions to M, rather Monte Carlo methods are used to sample the measure space randomly and thus to estimate S=Sfj-mj-fj log( fj/mj) (7) prob(D). If fj is the fractional flux in each cell, Sfj=1. mj incorporates Finally, the desired result, the mean spectrum, is given by any previous information available. Here there is none, so mj prob(M given D)=prob(D and M)/prob(D) (6) is a constant.Then, The prior1,8 prob(M) encapsulates the question that is to S=-Sfj log( fj/mj) (8) be asked of the data and forces recognition that the ultimate Calculating the entropy for each table yields S (positive quality of an answer depends as much on the question as on correlation)=0.28, S (negative correlation)=0.48 and S the data. The prior prob(M) attempts to identify that distri- (uncorrelated)=0.55, and this is in fact the most probable bution of the measure that best fits the data whilst making distribution given the data.the least commitment to missing data. Because no physical A recent paper by Sibisi and Skilling11 has shown that model is invoked in the prior, Bayesian analysis can cope with employing maximum entropy in the prior provides answers problems where the data are underspecified. that are dependent on the resolution with which the data are For example, the principle of maximum entropy can be recorded.They therefore proposed an alternative prior, invoked as a means of finding an optimum fit. Gull and employing the normalized power law, that avoids this diYculty Skilling9 have given an example of the principle of maximum and incorporated this into an algorithm termed Massive entropy in operation. An experiment is conducted to find out Inference (MI ). The MI algorithm (MaxEnt Solutions, how many kangaroos in Australia are both left handed and Cambridge, UK) was used for the calculations presented in have blue eyes.However, lack of forethought in the experimenthis paper. tal design results in two independent experiments being per- The benefits of this approach are as follows: formed that show that one third of the kangaroos are left (i) It is no longer necessary to begin with the assumption handed and one third have blue eyes. Therefore, the data are: that an absolute answer is provided by instrumental data and then to estimate after the event whether or not that assumption was valid.Rather, it is accepted that any Left-handed quantity inferred from a data set must be a probabilistic rather than an absolute quantity. This is the essence of True False the Bayesian approach. (ii) It can, uniquely, cope equally well with under- or over- Blue eyes True f1=x f2=1/3-x constrained data and responds by adjusting the pattern False f3=1/3-x f4=1/3+x of error bars accordingly. Where the data are overconstrained and the instrumentation provides a trans- There are an infinite number of solutions that can satisfy parent estimate of the measure, it will provide the same these criteria and some examples are given below.performance as conventional methods, but with a rigorous estimation of the errors. Positive correlation (all left-handed kangaroos have blue eyes): (iii) Only positive answers are possible and it follows that zeros in the spectrum become powerful constraints since all components that should contribute at that point are Left-handed unambiguously assigned a zero concentration.(iv) Knowledge is used formally in the PSF, and may be used True False iteratively to improve the fit between the observed and synthesized or ‘mock’ spectrum. Blue eyes True 1/3 0 False 0 2/3 (v) Errors are calculated independently for every data set (i.e., each sample) taking into account the uncertainties that lie therein. Negative correlation (no left-handed kangaroos have blue eyes): Experimental A synthetic mixture of rare earth elements was chosen as a Left-handed demonstration set because they are important elements in geochemistry and have well characterized elemental mass True False spectra in the region m/z 140–171 that exhibit isobaric interference and interference from oxides MO+ and hydroxides Blue eyes True 0 1/3 MOH+.False 1/3 1/3 The ICP-MS system used for these studies was a Model PQ2+ instrument (VG Elemental, Winsford, Cheshire, UK) run under standard operating conditions.A Model TR-30-C2 Uncorrelated: glass concentric nebuliser (J. E. Meinhard Associates, Santa Cruz, CA, USA), a Minipuls 2 peristaltic pump (Gilson Left-handed Medical Electronics, Villiers le Bel, France), and a VG Elemental water cooled spray chamber were used for sample True False introduction. The plasma was centred about the sampling orifice, which was in the region of 12 mm above the load coil.Blue eyes True 1/9 2/9 The sensitivity of the instrument thus configured was 70×106 False 2/9 4/9 counts s-1 ppm-1 for 115In. Data acquisition was carried out in the TRA mode using The best solution is intuitively that one which makes the PQ Vision software provided by VG Elemental. The TRA mode was chosen so that real time RSD values were available least commitment to the missing data. Another way of saying J. Anal. At. Spectrom., 1999, 14, 99–108 101for each mass channel (a required input for the spectral investigate the eVect of multiplicative interferences that could modify the PSF on a sample-to-sample basis. decomposition software).The acquisition parameters were 50 slices, 50 s data acquisition time (1 s per slice), 5000 ms delay The PSF employed here included the fixed isotopic abundances for each of the elements and therefore they are deter- time between points, 23 556 ms dwell time per point, 35 points (mass range m/z 140–176; 141 and 175 skipped; Pr, Lu, TbO mined as entities rather than individual isotopes.If isotope ratios are required, the variable isotopes must be included as not present in the spectrum), one point per isotope and a dead time of 33 ns. single species in the component menu. Stock standard solutions were 10 000 mg ml-1 Specpure solutions (Johnson Matthey, Royston, Hertfordshire, UK). High 2 Correlated measurements purity water used for dilution of the stock standard solutions was obtained by passing demineralised water through a labora- Mass spectrometry tory-reagent grade water system (Liquipure, Bicester, Fig. 2 shows a typical noise power spectrum derived from an Oxfordshire, UK) operated at 18 MV. All working solutions ICP source; in this case the signal was that from 80Ar2+. The were prepared so as to contain 2% v/v nitric acid of ultrapure plot shows the expected contributions of white noise and reagent grade. flicker noise. The latter comprises a 1/f component, pick-up Five test solutions were prepared, a blank plus four others at 50 Hz, source modulation at 100 Hz and higher harmonics containing 0.02, 0.5, 5 and 20 ppb each of Nd, Sm, Eu and and an audio component at about 330 Hz that is derived from Gd. Also, 5 ppm of Ba was added to each of these solutions the plasma dynamics. There is also a discrete signal component to provide a very high level of a potential interferent.at 3 Hz that is due to the rotation of the peristaltic pump, although this is not readily evident on this plot.These have been discussed elsewhere,12,13 but in any case the following Results and discussion observations can be made: Random or white noise is: Fig. 1(A) and (B) show comparisons of the synthesised spectra additive in nature; (mock spectra) with the actual spectra scaled to the number proportional to the square root of the signal magnitude; of standard deviations of the noise. The PSF has been gradually uncorrelated in diVerent measurement channels.modified to include the isotopic distributions of all the possible Multiplicative noise is: interferences. Adding BaO to the PSF largely resolves the proportional to (signal )y (y#1); poor fit at 154 u and in turn allows a more sensible distribution often correlated in diVerent measurement channels; of the observed counts to the isotopes of Sm (m/z 147, 148, correlated in adjacent regions of the ICP optical background 149, 150 152, 154).The 22s misfit at 155 u remains and also spectrum. the deviations at masses 156, 157, 158 and 160. Experimental The importance of the correlation is that when the signals data have shown that BaOH, not BaO, is the dominant oxide are measured on time-scales that are short compared with the species from Ba. Not placing this in the PSF has caused the characteristic period of the noise, subsequent ratioing or algorithm to try to account for the counts at 155 by allocating subtraction will lead to removal of the correlated noise. This them to 155Gd+, which then consequently over-estimates the is illustrated in Fig. 3(A), where two correlated isotopic signals contribution from the other isotopes of Gd (notably m/z 156, are represented schematically on a single time axis. To the left 157, 158 and 160). Once BaOH is included, the fit improves of the trace the vertical broken lines represent simultaneous dramatically. There are still problems at and above 159 u, but measurement of the signals whereas to the right the sloping by admitting the additional possibilities of NdO, SmO, EuO lines represent a small time displacement between the measure- and GdO [see Fig. 1(B)], the actual and fitted spectra can be ments in the two channels. The length of the lines represents made to agree within two standard deviations. the diVerence between the two signals that is constant for the Table 1 shows the output from the computations that have vertical lines, but of varying length for the sloping lines.eVectively provided the most probable distribution of the total To interpret this principle in terms of determining isotope counts between the species. Calibration for a particular element ratios by ICP-MS, consider Fig. 3(B), which represents a can be carried out by plotting the counts for that element typical measurement sequence in which three channels have against concentration in the usual fashion. For example, been allocated to each isotope.Each fixed period and their carrying out linear regression on the data for Nd, including combinations eVect a level of filtering on the signals.13 Of the blank as a sample, yields y=802x-19.96, standard error particular relevance to this discussion, however, is the time of the gradient=2.88, standard error of the intercept=26.6, lapse between measurement of the individual isotopes, which with r=0.999 98, i.e., a straight line, within error, passing is best estimated by the elapse time, given by through the origin. Although no Ce was added to the solutions, an unexpected background count of 1601 was observed at m/z elapse time=2tDnp+tS (9) 154; Ce was therefore included in the PSF as a possible contaminant.The computation has allocated a probable count where tD is the dwell time per channel, np is the number of of 154, but with a high level of uncertainty of ±530 counts, points per peak and ts is the settle time. For noise components that in turn leads to a low probability (8%) that Ce was with periods longer than twice the elapse time, isotope ratioing actually present.Additional masses at m/z 136 and 138 would will be beneficial in reducing noise. This is illustrated in have had to have been measured (in addition to m/z 140 and Fig. 3(C), which shows the improvement in measurement 142) to improve on this estimation. precision for the 107Ag-to-109Ag isotope ratio observed upon This exercise is indicative of how a Bayesian analysis can decreasing the elapse time, via the dwell time per channel, of be used to infer the interfering species that are determining which there were three per isotope, from 81.92 to 5.12 ms13 the actual data set.Alternatively, the algorithm can be given (the minimum practical settle time of 5.12 ms was employed). all possibilities at the outset and will return estimates with the The integration time per isotope was held constant at 49.152 s appropriate uncertainties. If some of these possible inter- by increasing the number of sweeps.Hence, reducing the ferences can be eliminated (e.g., from other information about elapse time from approximately 500 to 41 ms was eVective in the sample), the uncertainties reported from the computations reducing the detrimental influences of the correlated 1/f and can be reduced. The interferences encountered in this experi- pump induced noise. The results for Ag show that a fast sequential spectrometer ment are purely additive and more work is required to 102 J.Anal. At. Spectrom., 1999, 14, 99–10825 20 15 10 5 7 6 5 4 3 2 1 0 –1 –2 0 –5 –10 m/z A B No. of s's 140 142 144 146 148 150 152 154 156 158 160 162 164 166 170 158 160 172 174 176 164 162 166 168 170 172 174 176 154 156 152 150 148 146 144 142 140 Fig. 1 Plot of (measured ICP-MS data)-(data synthesized from Bayesian analysis) scaled to the standard deviation. The data are for 20 ppb each of Nd, Sm, Eu and Gd and 5 ppm of Ba in 2% nitric acid.(A) Three data sets are shown for the following assumptions concerning the species contributing to the observed spectra, Ce, Nd, Sm, Eu, Gd (black), add BaO (white), add BaO and BaOH (grey). (B) Two data sets are shown for the following assumptions concerning the species contributing to the observed spectra, Ce, Nd, Sm, Eu, Gd, BaO and BaOH (black), add NdO, SmO, EuO and GdO (white). Table 1 Output from Bayesian analysis of a mixed rare earth standard increases.Additionally, the eVects of mass bias have to be containing 20 ppb each of Nd, Sm, Eu and Gd plus 5 ppm of Ba taken into account if both accuracy and precision are of interest. Table 2 gives data for Pb acquired under optimum Element/ Probability measurement conditions on our venerable prototype VG PQ1 species (%) Counts ±Counts RSD (%) quadrupole ICP-MS instrument.14 The approximation to sim- Ce 8 154 530 344 ultaneous measurement was sustained by measuring each Nd 100 1601575 21042 1.32 isotopic pair separately.It was not found to be beneficial to Sm 100 1802738 28329 1.57 compromise this approximation by allocating more time to Eu 100 1808445 26540 1.47 less abundant isotopes, in this case 204Pb. Thallium was added Gd 100 1769622 20973 1.19 to the solutions to enable mass bias to be estimated from the BaO 100 443030 33750 7.62 205Tl-to-203Tl ratio using the power law approximation. BaOH 100 694991 30647 4.41 NdO 100 32318 3587 11.1 Operating parameters were rest mass=200 u, tJ=2 ms, tD= SmO 100 9928 1420 14.3 10.24 ms, three channels per peak, total acquisition time= EuO 18 321 707 220 120 s and detector dead time=32.5 ns.The data in Table 2 GdO 100 18220 1819 9.98 probably represent the limiting performance for sequential instruments. can emulate the measurement characteristics of a simultaneous Optical spectrometry instrument provided that the signals are acquired on appropriate time-scales.This is more diYcult when more than two Optical emission spectrometry has enjoyed a renaissance in recent times, almost entirely due to the advent of multi-channel isotopes have to be measured and when the mass range J. Anal. At. Spectrom., 1999, 14, 99–108 103Fig. 2 A typical noise power spectrum, as determined for 80Ar2, for an argon ICP coupled to the sampling interface of an ICP-MS instrument. 2D array detectors such as those based on CCD and CID technologies.15 Although simultaneous instruments with photomultiplier tube detection have been available for decades, these rely on local scanning of the spectrum to identify the background associated with each line.Array detectors, in contrast, provide true simultaneous measurement of line and adjacent background for either the whole spectrum or selected groups of lines for each element. Because, as stated earlier, noise in the background signal is correlated over limited wavelength ranges, when the oV-peak background signal is subtracted from the total signal at the analyte wavelength, the total variance is reduced by that component which is correlated in both channels, viz., s2=sa2+sb2-hsasb (10) where sa2 and sb2 are the variances at the peak and oV-peak wavelengths, respectively, and h is the degree of correlation having a value of +1 for full positive correlation, -1 for negative correlation and zero for uncorrelated signals. The dominant noise (for optical detection) close to the detection limit is that derived from the background and therefore optimizing the signal-to-background ratio is a simple and eVective means of achieving the lowest detection limits. Sadler et al.,16 however, have shown that when the background noise is correlated then optimization should be based on signal-to-square root of the background. This provided improvements in detection limits by factors ranging from 1 to 4.8.The actual improvement in precision and detection limit that can be achieved depends on a number of factors, including the degree to which the background signals are correlated and the magnitude of the background relative to the line intensity.The observed correlation will also vary depending on the number and location of the pixels allocated to line and background measurements and the structure of the back- Fig. 3 Schematic diagram of time correlated signals demonstrating the eVects of simultaneous measurement in two channels (vertical broken ground over the measurement window.Poussel and Mermet17 lines) and time displaced measurement (sloping broken lines). (B) derived an expression that relates the net relative standard Schematic timing diagram for the measurement of two isotopes by a deviation to the ratio of the actual concentration to the limit quadrupole mass filter operated in peak jump mode. TD=dwell time; of detection in terms of the degree of correlation and the TE=elapse time; TS=sweep or cycle time; TJ=setting time.(C) EVect flicker noise factor a: of dwell time on RSD for isotopic ratio measurement: & measured RSD for 107Ag5109Ag ratio; 2 counting statistic. RSDnet= Ca2+ 2(1-h)CL2 k2C2 D1/2 (11) 0.004. Below C/CL#20, there is an approximate order of magnitude improvement in RSD for h=0.99 compared with The latter is simply the ratio of the flicker noise magnitude to the net line signal. Fig. 4 shows data calculated from eqn. (11) h=0.This theoretical improvement was reflected in actual improvements in detection limits and limits of quantification (as for Fig. 1 in ref. 17) for h=0, 0.8, 0.95 and 0.99 with a= 104 J. Anal. At. Spectrom., 1999, 14, 99–108Table 2 Accuracy and precision of Pb isotope ratio measurement for that in the environmental and biological sciences, the transport, NIST SRM 981 pool dynamics and toxicology of the elements are dependent on their chemical form. Significant progress has been made by the direct coupling of separation techniques, notably the Sample/parameter 204Pb 206Pb 207Pb 206Pb 208Pb 206Pb various forms of HPLC, to powerful elemental detectors such as ICP-MS.18 The strength of such coupled techniques is that 1 0.0593 0.9153 2.1706 they provide low limits of detection, high accuracy and pre- 2 0.0591 0.9150 2.1715 cision and identification of elemental associations.However, 3 0.0592 0.9143 2.1715 this is achieved at the expense of losing molecular information, 4 0.0593 0.9146 2.1698 5 0.0593 0.9148 2.1712 particularly where organic ligands are involved.The study of Mean 0.0592 0.9148 2.1709 such ligands has traditionally been addressed by using, for RSD (%) 0.12 0.038 0.033 example, conventional forms of organic mass spectrometry. Counting statistic (%) 0.15 0.048 0.041 This is highly developed in terms of structural and qualitative Certified value 0.059042 0.91464 2.1681 analysis, but does not approach atomic spectrometric analysis 2s 0.000037 0.00033 0.0008 for accuracy and precision or, in some cases, power of detection.The challenge must be to bring the two approaches to bear simultaneously so that quantitative and qualitative information on both the inorganic and organic components can be obtained. 3.1 Separations Perversely, many problems in elemental speciation derive from the use of chromatography. Identification of species is by retention behaviour only; therefore, it is dependent on a priori information and the availability of suitable standards and there is no definitive identification of species.Only limited information is provided about the nature of the ligand and its bonding to the analyte. In many cases, where the complexes are thermodynamically and kinetically labile, they will tend to dissociate on the column, precluding the determination of stability constants. In order to achieve eYcient separation, column packings and solvent systems have to be tailored to specific forms of the analyte so that a complete analyte budget cannot be obtained in a single run.Retention on the column, ineYcient separation, dilution, contamination and incompatibility of solvent systems with the source add to the problems. Capillary electrophoresis. Recently, in this laboratory19 and elsewhere,20–23 capillary electrophoresis coupled to ICP-MS has been investigated as a technique that oVers an alternative to conventional chromatographic separations.The advantages of CE-ICP-MS can be summarized thus: it oVers high separation eYciencies of >3×105 equivalent theoretical plates; it 100 10 1 0.1 RSD (%) 10 100 C/ CL theta = 0 theta = 0.8 theta = 0.95 theta = 0.99 can handle positively, negatively and uncharged species in a Fig. 4 Improvement in RSD for simultaneous measurement of line single run; it can potentially handle labile complexes (e.g., and adjacent background intensities in ICP-AES.h=Degree of correnon- covalently bound) and colloidal systems; it is rapid com- lation; CL=limit of detection; a (flicker noise coeYcient)=0.004; k pared with, for example, gel permeation; it requires very small (confidence factor)=3. Data recalculated from ref. 17. samples of only a few nanolitres; and the columns are relatively simple and cheap. However, there are some disadvantages: by one order of magnitude under favourable conditions. compared with conventional chromatography, the detection However, above C/CL=20, the dominant noise source is limits are relatively poor; the resolution gain is significant only derived from the analyte line signal and therefore the improvefor high molecular mass compounds; it is diYcult to control ment in RSD is reduced, tending towards the limit (in this conditions and standardize migration behaviour; it is diYcult case 0.004) set by the flicker noise.A similar outcome will be to interface detectors; and the narrow capillaries tend to block, observed for lines in the low UV part of the spectrum where e.g., due to bacterial growth.the background intensity is lower and therefore less dominant Most serious of these is the relatively poor detection power. in determining the signal-to-noise ratio at low concentrations. For conventional on-column molecular absorption detection The above discussions illustrate the clear benefits, in terms this is a direct consequence of the very small optical pathlength.of gains in precision (and potentially in accuracy), aVorded For flow-through detectors such as the ICP, it derives from by simultaneous measurement systems when there is a substanthe necessity to transport the analyte to the detector. The only tial degree of correlation in the measured signals. It is perverse acceptable flow in a CE column is that due to electroen- that in analytical atomic spectrometry, such correlation is rare dosmosis, which varies from the order of 1 ml min-1 for and limited to specific situations.There does not appear to be uncoated capillaries to zero for coated capillaries. Hence a any prospect of finding universal reference signals. make-up flow is needed to transport the analytes that migrate from the capillary to the nebuliser and hence to the plasma. 3 Conserving molecular information In addition to analyte transport, the make-up flow performs two other important functions: it provides an electrical connec- Elemental speciation has been one of the principal growth areas in analytical atomic spectrometry.This reflects the fact tion to ground and it neutralizes the suction of the nebuliser J. Anal. At. Spectrom., 1999, 14, 99–108 105(if the nebuliser is self-aspirating), thus avoiding pressure 0.55 ml min-1. Data were acquired in the TRA mode and then processed in Excel 4 (Microsoft). induced flow in the column. Pressure induced flow has a laminar flow profile, compared with the ideal plug flow of electroendosmosis, and therefore produces considerable band Results and discussion.Fig. 6 shows a separation of Cr(III) and Cr(VI) run under conditions of 15 kV, 0.005 M borate broadening. Returning to the problem of low sensitivity, suppose 30 nl of sample are injected on to the column and buVer (pH 8.5) and make-up flow 0.01 M NH4Cl at 40 ml min-1. In this case the polarity was negative at the inlet then analytes migrate from the column with peak widths of 6 s.If a high eYciency nebuliser, running at 30 ml min-1 uptake end. The electroendosmotic flow was therefore towards the inlet, but this was compensated by residual suction from the rate, is used to transport the analyte to the plasma, the analyte is diluted in 6 s worth of nebuliser flow, i.e., 3 ml, a dilution nebuliser that produced a net bulk flow towards the detector. The expected order of appearance was therefore Cr(VI) fol- of 100-fold compared with the injected concentration.Clearly, it is important to transport as much of the aspirated analyte lowed by Cr(III), as shown. As indicated above, one of the potential advantages of CE as possible to the plasma. Currently, high eYciency nebulisers oVer transport eYciencies of the order of 30%. is that it can handle diVerent charge states of an element in a single run. This is indicated in Fig. 7, the data for which were acquired under the same conditions as those described for Cr.Experimental. Fig. 5 shows a schematic diagram of a prototype CE-ICP-MS interface constructed in our laboratory. Essentially, the CE capillary passed co-axially through a short length of electrically grounded Pt–Ir capillary (0.5 mm id) which in turn carried the make-up flow. Two types of nebuliser were used with this interface, a standard glass concentric nebuliser (Model TR-30-C2) and a high eYciency glass concentric nebuliser (Model HEN-170-AA) (both from J.E. Meinhard Associates). A laboratory-made U-shaped spray chamber of 28.5 ml internal volume was coupled to the nebulisers and the aerosol conducted to the plasma via 50 cm of 4 mm id flexible tubing. The ICP-MS instrument used was that described in Section 2. Subsequently an improved version of this interface has been developed.19 The CE system was built in-house. The capillary used was 100 mm id and 100 cm long (SGE, Milton Keynes, UK; ES2 coating). Power was derived from a Model 805 high voltage supply (Brandenburg, Croydon, Surrey, UK) capable of providing ±0–30 kV at 0–20 mA.The electrophoretic current was monitored with an ISO-Tech 90 Series portable multimeter Fig. 6 CE-ICP-MS separation of Cr(III ) and Cr(VI) in a (RS Components, Corby, Northants, UK) mounted in-line. 100 cm×100 mm id coated column. Conditions; separation voltage, 15 kV; 0.005 M borate buVer (pH 8.5); make-up flow, 0.01 M NH4Cl The high voltage components were mounted in a Perspex box at 40 ml min-1.with safety cut-oV switches that operated when the box was opened. The make-up flow was provided from a Minipuls 2 peristaltic pump (Gilson Medical Electronics) and capillary flushing was accomplished via a Model 341A syringe pump (Sage Instruments, Orion Research, Cambridge, MA, USA). Samples of 36 nl were injected by hydrostatic pressure diVerence. The optimum make-up flow for the high eYciency nebuliser was found to be of the order of 40 ml min-1.Indium and/or Ce were added to the make-up flow so that the stability and performance of the interface could be monitored. The high eYciency nebuliser running at 40 ml min-1 gave equivalent analytical figures of merit to the standard nebuliser run at Fig. 7 CE-ICP-MS separation of (a) 20 mg ml-1 Ni2+, (b) 20 mg ml-1 Ni2+ plus 5 mg ml-1 DMG, (c) 20 mg ml-1 Ni2+ plus 20 mg ml-1 humic acid and (d) 20 mg ml-1 Ni2+ plus 5 mg ml-1 DMG plus Fig. 5 Schematic diagram of a prototype CE-ICP-MS interface. 20 mg ml-1 humic acid. Detection as 60Ni. 106 J. Anal. At. Spectrom., 1999, 14, 99–108The lower trace (a) shows the signal obtained for 20 mg ml-1 unknown samples, prior separation is essential to permit sensible interpretation of the data. Currently, the ES and IS Ni2+ detected as 60Ni. Trace (b) shows the signal obtained when 5 mg ml-1 dimethylglyoxime (DMG) was added to the must be regarded as complimentary to other coupled elemental speciation techniques.This is well demonstrated in the work solution. The Ni–DMG complex is neutral and therefore acts as a bulk flow marker. Trace (c) shows the signal obtained of Corr and Larsen,28 who employed IS for the positive identification of As species previously separated by HPLC. from a mixture of 20 mg ml-1 Ni2+ and 20 mg ml-1 purified humic acid (Fluka, Gillingham, Dorset, UK). The broad peak Other examples are work by Pergantis et al.29 on As compounds and by Crews et al.30 on Se compounds.is due to the polymeric Ni–humic complex. Humic acid always gives broad peaks when separated by, e.g., size exclusion or CE because it is a polyelectrolyte containing a range of Low pressure/power plasma sources. Conventional atmosmolecular masses typically in the range 103–105 Da. The pheric pressure plasma sources running close to LTE, such as Ni–humic complex has a net negative charge and in this the ICP, have suYcient energy to eVect complete atomization configuration emerges ahead of the free Ni2+.Trace (d) was of samples introduced as aqueous or solid aerosols provided derived from a mixture containing 20 mg ml-1 Ni2+, 5 mg ml-1 that the particle diameter is less than about 5 mm. Sources DMG and 20 mg ml-1 humic acid. operating well away from LTE with low kinetic energy, such The importance of Fig. 7(d) is its significance to the study as the MIP, cannot cope with the thermal load associated with of trace metal bioavailability. Here are a free metal ion that the break-up of macroscopic particles. They do, however, have might be absorbed by a cell by a specific acceptor site, a high eVective excitation temperatures associated with the free neutral metal complex that would be lipophillic and therefore electrons and, depending on the applied power, these may be absorbable by a non-specific mechanism, and a high molecular present at suYcient density to produce complete fragmentation mass negatively charged complex that would not be bioavail- of molecular vapours. Moving to low pressures enables disable.This is an illustration of how CE has the potential to charges to be sustained at much lower energies because the provide a complete elemental budget. Such information could electrons are free to accelerate in the applied field without the also be used to quantify stability constants provided that damping eVect of collisions. Lowering the pressure too far retention in the capillary is minimised.causes a phase lag to develop (cosh�0, where h is the phase angle) between the applied field and the electron motion 3.2 Soft sources leading to poor energy transfer. This in turn yields insuYcient ionization to sustain the plasma. Under intermediate con- An alternative to employing separation techniques is to make ditions, a soft discharge can be sustained which, although a direct assault on the problem of speciation by using a source containing energetic electrons, has insuYcient density of high that preserves the molecular information.The sources energy particles to eVect molecular fragmentation. employed for organic mass spectrometry fall into this category, Recent work by Caruso and co-workers31,32 has sought to but they do not generally meet the requirements for accuracy, exploit such plasmas for speciation studies. Analytes must be precision and sample throughput of conventional elemental presented in vapour form and therefore GC has been used for analysis.Further, because of the high information content of prior conditioning and separation. A low pressure He ICP the signals, prior separation is often necessary to make sense running at 90W forward power was used to ionize and of the data. The sources investigated so far for elemental fragment organotin and organolead compounds and yielded speciation have been those which can be used in both molecular elemental detection limits in the range 1–14 pg. The same and elemental modes, which potentially simplifies data source running at 45W forward power yielded fragment ions interpretation and quantification. similar to those obtained in EI spectra.For perfluorotributylamine, the optimum power for the production of fragment Electrospray and ionspray. The electrospray (ES)24 and the ions was reduced to 5–8 W. Recently, Caruso’s group33 have gas assisted electrospray, sometimes referred to as the ionspray investigated a low power He MIP and have observed EI type (IS),25 are unique in being able to generate ions directly from spectra for perfluorotributylamine and tetramethyltin. Sample solution at atmospheric pressure for an energy expenditure of introduction was accomplished using a direct injection a few milliwatts. Because the ions produced may, under nebuliser.particular conditions, be the same as those present in solution, An alternative to electrodeless discharges is the rf assisted the technique has been considered to have potential for direct glow discharge, which can be run under both hard and soft speciation without recourse to prior separation.However, this conditions. The eVect of discharge operating parameters on claim has been challenged by Sharp and co-workers,26 who the fragmentation of organic compounds has been investigated commented on the complexity of the spectra that can be by Carazzato and Bertrand.34 Caruso and co-workers35 obtained from even simple single component salt solutions reported on the use of an rf-assisted glow discharge for the and on the fact that new species, not present in solution, are detection of organotin compounds separated by GC.observed in the mass spectrum. Further, the observed spectra Fragmentation patterns similar to those obtained by EI were are very sensitive to the operating conditions, which in turn observed and surprisingly these were relatively insensitive to vary from species to species.On the one hand it is very actual discharge operating parameters. convenient to have a source that is tunable and able to provide intact molecular ions or bare elemental ions, but choosing the observed species is somewhat removed from the concept of an Conclusion ideal transparent detector. Quantification is diYcult, requiring the use of internal standards for calibration. Detection limits This paper has attempted to demonstrate that atomic spectrometric methods of analysis are capable of providing a great are also much poorer (ng ml-1 range) than those obtained with conventional atomic spectrometry sources.A practical wealth of information on chemical systems. The future challenge must be to find means of combining the virtues of atomic diYculty associated with many low flow techniques (typical flow rates are 5–10 ml min-1) is cross-contamination of the spectrometry with the qualitative and structural analysis capabilities of other analytical techniques so that atomic and source unit, which mitigates against high sample throughput.Electrospray and ionspray are undoubtedly powerful tools molecular information can be obtained simultaneously without loss of detection power, accuracy, precision or sample for studying speciation in the solution phase when there is prior knowledge of the components.27 However, to address throughput. J. Anal. At. Spectrom., 1999, 14, 99–108 10719 K. A.Taylor, B. L. Sharp, D. J. Lewis and H. M. Crews, J. Anal. References At. Spectrom., 1998, 13, 1095. 20 J. W. Olesik, J. A. Kinzer and S. V. Olesik, Anal. Chem., 1995, 1 Maximum Entropy and Bayesian Methods, ed. J. Skilling, 67, 1. Cambridge University Press, Cambridge, 1988, and Kluwer, 21 Q. H Lu and R. M. Barnes, Microchem. J., 1996, 54, 129. Dordrecht, 1989. 22 K. Sutton, R. M. C. Sutton anJ. A. Caruso, J. Chromatogr. A, 2 H. Martens and T. Naes, Multivariate Calibration, Wiley, New 1997, 789, 85. York, 1989. 23 B. Michalke and P. Schramel, Electrophoresis, 1998, 19, 270. 3 M. E. Ketterer and D. A. Biddle, Anal. Chem., 1992, 64, 1819. 24 J. D. Fenn, M. Mann, C. K. Meng and S. F. Wong, Mass 4 A. Behrens, Spectrochim. Acta, Part B, 1995, 50, 1521. Spectrom. Rev., 1990, 9, 37. 5 A. Donachie, A. D. Walmsley and S. J. Haswell, Anal. Commun., 25 E. C. Huang and J. D. Henion, J. Am. Soc. Mass Spectrom., 1990, 1996, 33, 293. 1, 158. 6 D. A. Sadler and D. Littlejohn, J. Anal. At. Spectrom., 1996, 26 B. L. Sharp, A. B. Sulaiman, K. A. Taylor and B. N. Green, 11, 1105. J. Anal. At. Spectrom., 1997, 12, 603. 7 M. R. Cave, J. Anal. At. Spectrom., 1998, 13, 125. 27 I. L. Stewart and G. Horlick, J. Anal. At .Spectrom., 1996, 11, 8 G. J. Daniell, in Maximum Entropy in Action, ed. B. Buck and 1203. V. A. Macaulay, Clarendon Press, Oxford, 1981. 28 J. J. Corr and E. H. Larsen, J. Anal. At. Spectrom., 1996, 11, 1215. 9 S. F. Gull and J. Skilling, IEE Proc., F: Radar Signal Process., 29 S. A. Pergantis, W. Winnik and D. Betowski, J. Anal. At. 1984, 131, 646. Spectrom., 1997, 12, 531. 10 E. T. Jaynes, Proc. IEEE, 1982, 70, 939. 30 H. M. Crews, P. A. Clarke, D. J. Lewis, L. M. Owen, P. R. Strutt 11 S. Sibisi and J. Skilling, J. R. Statist. Soc. B, 1997, 59, 217. and A. Izquierdo, J. Anal. At. Spectrom., 1996, 11, 1177. 12 M. P. Goudzwaard and M. T. C. de Loos-Vollebregt, Spectrochim. 31 T. M. Castillano, J. J. Giglio, E. H. Evans and J. A. Caruso, Acta, Part B, 1990, 45, 887. J. Anal. At. Spectrom., 1994, 9, 1335. 13 I. S. Begley and B. L. Sharp, J. Anal. At. Spectrom., 1995, 10, 171. 32 G. O’Connor, L. Ebdon, E. H. Evans, H. Ding, L. K. Olson and 14 I. S. Begley and B. L. Sharp, J. Anal. At. Spectrom., 1997, 12, 395. J. A. Caruso, J. Anal. At. Spectrom., 1996, 11, No. 12, 1151. 15 T. W. Barnard, M. I. Crockett, J. C. Ivaldi, P.L. Lundberg, D. A. 33 N. P. Vela, J. A. Caruso and R. D. Satzger, Appl. Spectrosc., 1997, Yates, P. A. Levine and D. J. Sauer, Anal. Chem., 1993, 65, 1231. 51, 1500. 16 D. A. Sadler, D. Littlejohn and C. V. Perkins, J. Anal. At. 34 D. Carazzato and M. J. Bertrand, J. Am. Soc. Mass Spectrom., Spectrom., 1996, 11, 463. 1994, 5, 305. 35 L. K. Olson, M. Belkin and J. A. Caruso, J. Anal. At. Spectrom., 17 E. Poussel and J. M. Mermet, Spectrochim. Acta, Part B, 1996, 1996, 11, 491. 51, 75. 18 S. J. Hill, M. J. Bloxham and P. J. Worsfold, J. Anal. At. Spectrom., 1993, 8, 499. Paper 8/07472K 108 J. Anal. At. Spectrom., 1999, 14, 99–108
ISSN:0267-9477
DOI:10.1039/a807472k
出版商:RSC
年代:1999
数据来源: RSC
|
2. |
Determination of total selenium in serum, whole blood and erythrocytes by ICP-MS |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 109-112
Christine E. Sieniawska,
Preview
|
|
摘要:
Determination of total selenium in serum, whole blood and erythrocytes by ICP-MS† Christine E. Sieniawska,* Ruth Mensikov and H. Trevor Delves SAS Trace Element Unit, Chemical Pathology, Southampton General Hospital, Southampton, UK SO16 6YD Received 10th August 1998, Accepted 3rd November 1998 A simple method for measuring selenium in serum by ICP-MS using a 1+14 dilution of a 100 ml sample in a diluent containing 1.0% v/v butan-1-ol to eliminate interferences at 78Se from argon adduct ions has been shown to be accurate and precise in a 24 month evaluation for routine clinical assays and has been developed to allow analysis of whole blood and erythrocytes.Excellent performances were achieved in two External Quality Control Assessment programmes for the measurement of selenium in serum: Centre du Toxicologie de Quebec and Trace Element Quality Assessment Scheme from the University of Surrey (Guildford). The regression equation for our ICP-MS data versus target values for analysis of Se in 109 sera from November 1996 to March 1998 was: ICP-MS=0.969x Target+0.031 mmol l-l, r=0.996. For serum analysis the detection limit was 0.02 mmol l-1, and the RSDs at 0.3–2.0 mmol l-1 were 2.0–4.0% within-run and 4.3–10.0% between-run.Initial applications of this method to measuring selenium in whole blood (1+14 dilution) and red blood cells (1+29 dilution) gave detection limits of 0.02 mmol l-1; RSDs of 2.1–5.8% within-run, and of 5.0–8.1% between-run, at concentrations of 0.3–2.0 mmol l-1.Analysis of a whole blood reference sample (Seronorm 404107), with a target concentration of 1.01 mmol l-1 and acceptable range of 1.01–1.15 mmol l-1, gave 1.06±0.06 mmol l-1 (mean±s). Table 1 Isobaric interferents on selenium determination by ICP-MS Introduction Isotope Abundance (%) Isobaric interferent A method for measuring selenium in serum by ICP-MS using a 1+14 dilution of a 100 ml sample in a diluent containing 74Se 0.9 38Ar36Ar 74Ge 1.0% v/v butan-1-ol to eliminate interferences at 78Se from 76Se 9.4 40Ar36Ar 38Ar2 76Ge argon adduct ions previously reported as a communication1 is 77Se 7.6 40Ar37Cl presented following a two year evaluation for routine clinical 78Se 23.8 40Ar38Ar 78Kr 80Se 49.6 40Ar2 assays. 82Se 8.7 40Ar2H2 82Kr 40Ar42Ca 81BrH Serum and whole blood selenium reflect nutritional intake, and have proved to be useful indicators of selenium depletion. Functional deficiency, however, has been assessed by monitorcaused by the formation of polyatomic ions in the plasma.ing the activity of glutathione peroxidase.2 This selenoprotein Although selenium has six isotopes (Table 1), there are a enzyme, present mainly in the erythrocytes, protects memnumber of argon adducts originating from the plasma support branes from damage caused by the peroxidation of lipids. gas, which themselves and in combination with other ions Selenium levels in erythrocytes would therefore be a more from the sample matrix, show the same nominal mass as these eVective indicator of selenium status.For this reason, a precise isotopes, causing false positive results. All of the selenium and accurate method of selenium analysis in whole blood and isotopes are aVected as can be seen by scanning a blank red blood cells is necessary. acidified with hydrochloric acid (Fig. 1). Hydride generation coupled to atomic absorption Therefore accurate measurement of selenium by ICP-MS spectroscopy3 is widely used for the measurement of selenium can only be achieved by the separation of the isotopes from in whole blood, but given the need to acid-digest the sample these isobaric ions (e.g., using a high resolution magnetic before analysis, the sample preparation is time consuming.sector field instrument or by coupling to a hydride generation Direct analysis by electrothermal atomic absorption specsystem), 5 or by reducing the level of these interferences.The troscopy is possible, but requires the careful use of chemical significant reduction of the concentration of argon adducts by modifiers in order to stabilise the selenium at the high temperature necessary for the removal of matrix components, and additional problems are caused by the spectral interference due to the high concentration of iron in the blood.4 In recent years, inductively coupled plasma mass spectrometry (ICP-MS) has proved to be one of the most versatile techniques for trace metal analysis, lending itself to direct measurement of a number of analytes in various biological matrices.However, for selenium, analysis by low resolution ICP-MS has been hampered by spectroscopic interferences †Presented at the Ninth Biennial National Atomic Spectroscopy Fig. 1 Isobaric interferents on selenium determination by ICP-MS. Symposium (BNASS), Bath, UK, July 8–10, 1998. J. Anal. At. Spectrom., 1999, 14, 109–112 109Table 2 Operating conditions for the ICP-MS 1% ammonia, 1.16 g l-1 EDTA, in 0.33% ammonium phosphate. Addition of these diluents allows the viscosity of the Instrument SCIEX Elan 5000 sample to be reduced, preventing the blockage of the torch Rf power 1.0 kW injector tube during the analytical runs.Indium was added as Plasma gas flow rate 15.00 l min-1 an internal standard, and the solutions were made up to 3 ml Nebuliser gas flow rate 1.05 l min-1 Intermediate gas flow rate 1.00 l min-1 with butanol and de-ionised water.The eVect of increasing Sampling cone Nickel, 0.75 mm orifice diameter butanol concentrations on diluted sera was investigated. Skimmer cone Nickel, 0.75 mm orifice diameter Torch Standard demountable quartz torch with 2.0 mm id alumina injector tube Results Sample uptake rate 0.8 ml min-1 (by peristaltic pump) Establishing the optimum butan-1-ol concentration Nebuliser Cross-flow Measurements 80 ms dwell time, 75 sweeps per reading, As the butanol was increased up to 4% v/v, the signals at 1 reading per replicate, 2 replicates masses 76, 77, 78, and 82, from aqueous blanks, decreased to Masses 76, 77, 78, 82 and 115 a maximum of threefold (Fig. 2). On introduction of a serum sample, all the selenium isotopes gave the greatest signal at the introduction into the plasma of either a molecular gas, 1% butanol. such as nitrogen, or a simple organic molecule such as propan- Matrix-matched calibrating standards were prepared by 2-ol, was first described by Evans and Ebdon.6 It is thought diluting equal volumes of a series of aqueous selenium stanthat under such conditions, through charge-transfer mechan- dards containing selenium up to 200 mg l-1 in 0.1% v/v nitric isms, the argon adducts become involved in the formation of acid.Aqueous, bovine serum and bovine blood calibration carbides, nitrides and oxides, eVectively shifting the mass of curves can be seen in Fig. 3 for the selenium isotopes at masses the polyatomic ions away from the region of interest.Selenium 77, 78 and 82, the matrix suppression of the signals confirming has successfully been determined in serum using ethanol,7 and the need to use matrix-matched standards in the analysis of methanol,8 and based on this approach of diluting the sample serum and whole blood specimens. with an organic molecule, we reported the measurement of Analysis of a certified quality control serum (Seronorm selenium in serum using butan-1-ol.1 The evaluation of this 311089, Nycomed Pharma Diagnostics, Oslo, Norway), and method for routine analysis of serum/plasma for selenium for External Quality Control Assessment (EQA) sera with estabtwo years and its development to allow analysis of whole lished selenium concentrations, showed good agreement blood and erythrocytes are reported here.between the target concentrations and the observed results for the three isotopes. Our own internal quality control material, consisting of spiked bovine sera, matched the target values Experimental established by hydride generation ETA-AAS, for isotopes 77Se Instrumentation and 78Se, but were higher when determined by 82Se, suggesting the presence of interferences.On scanning the two types of A Perkin-Elmer SCIEX Elan 5000 ICP-MS was used, with the sera, it was seen that the bovine serum contained a considerably operating conditions given in Table 2. greater level of bromine, indicating that the 81Br ion may have combined with hydrogen, to give hydrogen bromide at mass Reagents 82 (Fig. 4). Matrix-matched calibration standards were prepared using The analytical data reported here, and in our previous bovine serum and whole blood (Selbourne Biological Services, paper,1 indicate that 78Se is the appropriate isotope for sel- Alton, Hampshire, UK), with a stock selenium solution, at enium determination by this method. For routine samples, the 1020 mg l-1 (Sigma-Aldrich, Poole, Dorset, UK; Merck, analysis had been scaled down so that 100 ml sample aliquots Poole, Dorset, UK).The samples were diluted with butan- could be used per replicate, and measurement of only 78Se 1-ol, indium solution from a stock solution at 10 000 mg l-1, and 115In reduced the analysis time to less than 2.5 min per Triton-X-100, ammonia, (NH4)2H2EDTA, and ammonium sample. Selenium analysis in plasma samples showed similar phosphate, all of ‘AristaR’ or ‘Spectroscol’ purity (Merck).matrix eVects to bovine serum (Fig. 5), and therefore could Distilled, then de-ionised water (Milli-Q system, Millipore, be assayed against a bovine serum calibration curve. Watford, Hertfordshire, UK) and ‘AristaR’ nitric acid Precision and accuracy of selenium measurement in serum (16 mol l-1, Merck) were used throughout. The detection limit, three times the standard deviation of Preparation of samples replicate blank measurements, was less than 0.02 mmol l-1.The precision of this method was found to vary from 9% 200 ml Bovine serum samples were diluted with equal volumes of 1% Triton-X-100, and a modifier solution which contained standard deviation at 0.25 mmol l-1 to half that at Fig. 2 Change in signal intensities for Se isotopes on increasing the butanol concentration; for aqueous blanks (a) and a spiked bovine serum sample (b). 110 J. Anal. At. Spectrom., 1999, 14, 109–112Fig. 3 Aqueous, bovine serum and whole blood calibration curves: (a) 77Se; (b) 78Se; and (c) 82Se. 1.8 mmol l-1, between runs (Fig. 6). Our performance in two External Quality Control schemes, during a period of one and a quarter years, showed good agreement between reported results and the target concentrations (Fig. 7). The diVerences between the two values are individually plotted in Fig. 8. Fig. 4 Mass scans of bovine serum (a) used in the preparation of Fig. 6 Precision of serum selenium determination by ICP-MS. internal quality control samples and of human serum (b) showing the diVerence in bromine content.Fig. 7 Performance in EQA schemes for selenium in serum; proximity Fig. 5 Regression lines for 78Se measurement in bovine serum and plasma. to target values. (November 1996 to February 1998.) J. Anal. At. Spectrom., 1999, 14, 109–112 111Table 3 Internal quality control data for selenium in whole blood Target value/ Observed mean/ RSD (%) within-run RSD (%) within-run RSD (%) between-run mmol l-1 mmol l-1 (serum method) (modified method) (modified method) QcA 0.37±0.03 5.8 3.4 8.1 Bovine blood(1+1) n=10 n=10 n=92 QcB 0.80±0.04 2.9 2.9 5.0 Bovine blood n=10 n=10 n=92 Seronorm level 1 1.01 1.06±0.08 3.0 2.1 7.6 (404107) (1.01–1.15) n=10 n=10 n=91 Seronorm level 2 1.04 1.06±0.08 7.6 (404108) (1.04–1.13) n=75 aqueous blanks had been increased by 20%, the increase in sample5noise signal ratio was greater, giving better precision. The parameters on the ICP-MS were changed so that 100 sweeps per reading were carried out on analysis of each sample; the slightly longer analysis time increased the precision of the method, particularly at low selenium levels, as can be seen by the improved within-run %RSDs of the internal quality control material (Table 3).Measurement of selenium in erythrocytes Selenium measurement in erythrocytes was carried out on a 1+1 dilution of the red blood cells, with our ammonia–phosphate–EDTA diluent.The matrix eVect of the diluted erythrocytes was similar to that of the bovine blood, as can be seen by the parallel calibration curves in Fig. 9. The samples were analysed against a bovine blood calibration curve, and in the absence of erythrocyte quality control material, whole blood Seronorm, levels 1 (404107) and 2 (404108), were used, as well as the endogenous selenium level in the bovine blood, the bovine blood being diluted 1+1 (Table 3). The measured erythrocyte selenium levels were expressed in Fig. 8 DiVerences between reported values and target values. (Zone terms of mmol per haemoglobin concentration for each sample, parameters are defined by the TEQAS scheme.) in order to compensate for any dilution caused by the presence of plasma in the specimen. Preliminary analysis shows that the selenium levels in erythrocytes of healthy subjects are on average 1.6 times that found in the plasma for each individual. Conclusion The method described here is suitable for routine analysis of selenium, requiring only 100 ml of sample per replicate, very little sample preparation, only a simple dilution of the specimen, and a short analysis time of 140 s for serum and 160 s for each blood solution. The work carried out in this paper was supported by the Department of Trade and Industry as part of the National Measurement System Valid Analytical Measurement Programme.References 1 H. T. Delves and C. Sieniawska, J. Anal. At. Spectrom, 1997, 12, 387. 2 D. McMaster, N. Bell, P. Anderson and A. Love, Clin. Chem., Fig. 9 Regression lines shown for the addition calibration of 78Se in 1990, 36, 211. whole blood and red blood cells. 3 O. Mestek, M. Suchanek, Z. Vodickova, B. Zemanova and T. Zý�ma, J. Anal. At. Spectrom., 1997, 12, 85. 4 B. Radziuk and Y. Thomassen, J. Anal. At. Spectrom., 1992, 7, 397. Measurement of selenium in whole blood 5 M. P. Rayman, F. R. Abou-Shakra and N. I. Ward, J. Anal. At. Spectrom., 1996, 11, 61. Measurement of whole blood selenium was found to require 6 E. H. Evans and L. Ebdon, J. Anal. At. Spectrom., 1989, 4, 299. a number of modifications to the serum method. The indium 7 J. Goossens, F. Vanhacke, L. Moens and R. Dams, Anal. Chim. spike was increased 2.5 fold to compensate for the greater Acta., 1993, 280, 137. suppression by the matrix of analyte sensitivity. The rf power 8 R.Mun� oz Olivas, C. R. Que�tel and O. F. X. Donard, J. Anal. At. was increased to 1.05 kW, in order to decompose eVectively Spectrom., 1995, 10, 865. the greater solids content, of around 220 g l-1 protein in blood as opposed to 70 g l-1 in plasma. Although the 78Se signal of Paper 8/06307I 112 J. Anal. At. Spectrom., 1999, 14, 109&nda
ISSN:0267-9477
DOI:10.1039/a806307i
出版商:RSC
年代:1999
数据来源: RSC
|
3. |
Determination of cadmium in biological and environmental materials by isotope dilution inductively coupled plasma mass spectrometry: effect of flow sample introduction methods |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 113-120
J. Pablo Valles Mota,
Preview
|
|
摘要:
Determination of cadmium in biological and environmental materials by isotope dilution inductively coupled plasma mass spectrometry: eVect of flow sample introduction methods† J. Pablo Valles Mota, M. Rosario Ferna�ndez de la Campa, J. Ignacio Garcý�a Alonso* and Alfredo Sanz-Medel Department of Physical and Analytical Chemistry, University of Oviedo, 33006 Oviedo, Spain Received 21st September 1998, Accepted 30th November 1998 Experimental parameters governing the instrumental precision and accuracy for isotope ratio measurements of cadmium in ICP-MS, including sampling time, mass bias, detector dead-time and spectroscopic interferences, were first characterised.The precision achieved for the isotope ratio determination of Cd was around 0.2% (RSD for n=5) both for 1115112 and 1115114 ratios. Two alternative flow approaches for the determination of ultratrace concentrations of cadmium by isotope dilution (ID) were explored and compared with the more conventional ID methodology: first, on-line mixing of the sample solution with the spike solution just before the ICP-MS nebuliser using a peristaltic pump and, second, the generation of volatile cadmium species using sodium tetraethylborate by merging zones flow injection ICP-MS. The three approaches were successfully applied to the determination of ultratrace levels of cadmium in biological and environmental certified reference materials (NIST SRM 2670 Freeze- Dried Urine, IAEA H-8 Horse Kidney, BCR TP-25 Lichens, PACS-1 Marine Sediment and SLRS-3 Riverine Water).The on-line ID method proved to be the most convenient for the determination of cadmium in such samples because it is fast, provides similar results to those of conventional ID and requires less sample preparation. discrimination and detector dead-time systematic errors have Introduction to be examined taking into account random error propaga- Cadmium is a highly toxic metal and therefore it must be tion.9 frequently monitored in biological and enviromental samples.1 In recent years, the use of ID-ICP-MS for the determi- Atomic spectrometry, particularly electrothermal atomic nation of cadmium in biological and environmental samples absorption spectrometry, has been widely used for this has increased dramatically. The determination of cadmium purpose.Unfortunately, the determination of ultratrace by ID-ICP-MS in sediments,10–12 riverine water,11,13,14 seaamounts of cadmium in some important high salinity samples water,14,15 atmospheric particulates,16 Antarctic snow and (e.g., sea-water and human biological fluids) is not an ice,17 urine,11,14,18–20 animal tissues18,19 and rice.21 In many easy task, owing to the high background caused by the cases, acid dissolution aided by microwave power was applied matrix.to the solid samples.10,11,14,18,19,21 Furthermore, most of these Inductively coupled plasma mass spectrometry (ICP-MS) studies used previous several separation and preconcentration is a sensitive and precise technique for trace element steps: solvent extraction with dithizone;16 chelation by silicadeterminations and has been satisfactorily applied to the inmobilized 8-hydroxyquinoline;13,21 preconcentration by an determination of cadmium in diverse samples.2–6 SO3-carboxymethyl-quinolin-8-ol-cellulose column;14 copre- Furthermore, it permits very rapid isotope ratio determi- cipitation with Mg(OH)2;15 anodic stripping voltammetry nations, allowing for stable isotope dilution techniques.using a reticulate vitreous carbon working electrode to Isotope dilution (ID) analysis is a powerful strategy capable deposit and separate the analyte;18 vapour generation with of highly accurate and precise determinations of elements NaBH4;11 and electrothermal vaporization.19,20 that have at least two stable isotopes.7 The ID technique In this work two new flow approaches for ID determination compares the natural abundance of two isotopes of an of cadmium were explored and compared with the more element with the abundance occurring in the sample after an conventional ID methodology. First, the on-line ID enriched isotope solution (spike) has been added.Since method,22–25 achieved by merging the sample solution with another isotope of the same element represents the ideal the spike solution just before the ICP-MS nebulizer using a internal standard for that element, ID results are expected to peristaltic pump, which have never been applied until now to be highly accurate and precise even when the sample contains cadmium determination, was applied. Second, the generation high concentrations of concomitant elements and/or there is of volatile cadmium species using sodium tetraethylborate, sample loss during the preparation or pre-treatment processes.previously studied in our laboratory as an eYcient gaseous According to some workers, ID-MS may already qualify as a cadmium introduction method,26 is reported for the first time definitive method, under certain conditions, in the certification for ID analysis and was studied using merging zones flow of reference materials.8 However, for ICP-MS isotopic injection ICP-MS.The analytical performance characteristics applications, the eVects of isobaric interferences, mass of the corresponding ID methods were investigated in detail and compared for the determination of ultratrace levels of cadmium in biological and environmental reference †Presented at the Ninth Biennial National Atomic Spectroscopy Symposium (BNASS), Bath, UK, July 8–10, 1998.materials. J. Anal. At. Spectrom., 1999, 14, 113–120 113a two-channel peristaltic pump (Minipuls 2, Gilson, Experimental Worthington, OH, USA), two six-way valves (Omnifit, Instrumentation Cambridge, UK) and a gas–liquid separator, described previously27 as shown in Fig. 1(c). All FI coils in the manifold The inductively coupled plasma mass spectrometer used was were made of 0.8 mm id Teflon tubing.a Hewlett-Packard Model HP4500 (Yokogawa Analytical Systems, Tokyo, Japan). The operating conditions were optim- Reagents ized for the ID determination of cadmium (Table 1). A Milestone (Socisole, Italy) Model 1200 microwave digester A 1000 mg l-1 cadmium(II ) stock standard solution stabilized with a EM-45/A extractor module and a AC-100 open/close in 0.5 M HNO3 was purchased from Merck (Darmstadt, module with medium-pressure PTFE vessels were employed Germany).Enriched 111Cd was obtained from Cambridge for the digestions of the samples. Isotope Laboratories (Woburn, MA, USA) as CdO and Diagrams of the three experimental devices used are shown dissolved in high purity nitric acid (Suprapur, Merck). The in Fig. 1. concentration of this solution was established by reverse The merging zones flow injection manifold consisted of isotope dilution analysis.Table 2 gives the isotope abundances for natural Cd28 and the measured isotopic composition of Table 1 ICP mass spectrometer operating and data acquisition the enriched 111Cd solution, which was not certified for isotope parameters composition. Sodium tetraethylborate(III ) solutions were prepared by Operating parameters— dissolving the solid reagent (Strem Chemicals, Strasbourg, Rf power 1300 W France) in ultrapure water, stabilized in 0.1% m/v sodium Plasma gas flow rate 15 l min-1 Auxiliary gas flow rate 1.0 l min-1 hydroxide solution, and stored in the dark at 4 °C in polypropy- Nebulizer gas flow rate 1.3 l min-1 lene vessels.Aqueous solutions were stable in capped bottles Sample uptake rate 0.4 ml min-1 for about 2 weeks when stored at 4 °C in the dark. Sampling depth 5.8 mm All other reagents were of analytical-reagent grade, and all Sampler and skimmer Ni; 1 and 0.4 mm id dilutions were prepared with filtered (0.22 mm) 18 MV Nebulizer type Babington de-ionized water provided by a Milli-Q system (Millipore, Spray chamber Scott type, double pass (temperature 2 °C) Milford, MA, USA).Detector Secondary electron multiplier Environmental and biological materials analysed in this work included NIST SRM 2670 Freeze-Dried Urine from Data acquisition parameters— the National Institute of Standards and Technology Points per peak 1 (Gaitherg, MD, USA), IAEA H-8 Horse Kidney from Scans per point 1000 the International Atomic Energy Agency ( Vienna, Austria), Integration time 10 s per point Dwell time 10 ms SLRS-3 Riverine Water and PACS-1 Marine Sediment from Replicates 5 the National Research Council (Ottawa, Canada) and BCR TP-25 Lichens from the Community Bureau of Reference (Brussels, Belgium).Solid and digested samples were conserved at 4 °C in the dark. Sample preparation PACS-1 Marine Sediment. A portion of 0.15 g sample was weighed directly into a PTFE vessel and 3 ml of HNO3 (65% m/v) and 1 ml of HF (48% m/v) were added.After tightly capping the vessels, the sample carousel was placed in the microwave oven and the following heating programme was started: (1) 450 W, 2 min; (2) 600 W, 6 min; (3) 450 W, 2 min. After cooling, 1 ml of H2O2 (30% m/v) was added to the samples and the digestion heating programme was repeated. Finally, 1 ml of H3BO3 (5% m/v) was added to complex the excess HF and the vessels were heated in the microwave oven at 450 W power.The solutions were transferred into polypropylene containers and diluted by mass to about 50 g with ultrapure water. Table 2 Isotopic abundances (at.-%) of natural cadmium28 and the 111Cd spike (measured) Mass Natural28 Spikea 106 1.25 0.01 108 0.89 0.01 110 12.49 0.63 111 12.80 96.25 112 24.13 1.82 113 12.22 0.43 114 28.73 0.61 116 7.49 0.23 aDetermined after mass discrimination correction using natural Fig. 1 Schematic diagrams of the three ID-ICP-MS methods cadmium.compared. 114 J. Anal. At. Spectrom., 1999, 14, 113–120IAEA H-8 Horse Kidney. Amounts of 0.1 g of sample were When the sample is mixed on-line with the spike, as Fig. 1(b) shows, the equation used was weighed into PTFE vessels and 4 ml of HNO3 (65% m/v) were added. The vessels were heated in the microwave oven using the two step programme (1) 300 W, 2 min and (2) 600 W, Cs=Cst dst ds Rs-Rst Rs-Rm Rsp-Rm Rsp-Rst 4 min. This resulted in a clear digest. After cooling, the samples were diluted to about 50 g with ultrapure water.where Cs is the unknown concentration of cadmium in the sample (s) and Cst is the concentration of cadmium in the BCR TP-25 Lichens. A portion of 0.2 g of sample was natural standard (st), dst and ds are the density of the standard weighed into a PTFE vessel and 3 ml of HNO3 (65% m/v), and the sample, respectively, Rst and Rm are the measured 1 ml of HClO4 (74% m/v) and 1 ml of HF (48% m/v) were 114Cd/111Cd isotope ratios for the mixtures of the spike with added.The vessels were placed in the microwave oven and the the standard and the sample, respectively, and Rsp and Rs are same heating programme as used in the PACS-1 digestion the same ratios in the sample and the spike, respectively. procedure was started. After cooling, 1 ml of H3BO3 (5% m/v) The same equations were used for 112Cd as reference isotope was added and the samples were heated for 4 min at 450 W. instead of 114Cd.Finally, the samples were diluted to about 50 g with ultrapure water. Results and discussion NIST SRM 2670 Freeze-Dried Urine. This reference material Optimization of instrumental parameters was reconstitued as described in the certificate and analysed Typical plasma operating conditions in the ICP-MS are sum- directly after appropriate dilution with ultrapure water. marized in Table 1. The monitored masses were 111, 112 and 114 for sample analysis and other Cd isotopes (106, 108, 110, Procedures 113 and 116) were also monitored in preliminary experiments General liquid on-line ID-ICP-MS.Using the two-channel on the accuracy and precision of isotope ratio measurements peristaltic pump of the ICP-MS instrument, a 111Cd spike with the instrument. The ion lens voltage settings and other solution is continously pumped by one channel whereas the parameters of the instrument were tuned daily with a 25 ng g-1 samples or natural Cd standards are sequentally pumped by Cd solution to obtain the maximum signal.It was observed the other channel. The two flows merge in a T-piece where the that some experimental conditions used aVected not only the sample or the standard is mixed with the spike. The resulting signal obtained for the Cd isotopes but also the value and the mixture is then introduced into the nebulizer of the ICP-MS precision of the measured isotope ratios. This was clearly instrument at a flow rate of 0.4 ml min-1.observed during the optimization of the ion lenses. As an example, Figs. 2 and 3 show the values obtained for the 114Cd General ID-FI-vapour generation-ICP-MS. Using the flow and the 111Cd/114Cd ratio while optimising the Einzel 1 and 3 system shown in Fig. 1(c), 0.04 M HNO3 and water are con- and the Omega (+) lenses, respectively. As can be observed, tinuously pumped through the two channels of a peristaltic on applying diVerent voltages to the Einzel 1 and 3 lenses pump at a flow rate of 3 ml min-1.Volumes of 300 ml of (Fig. 2) there is a continuous change in the isotope ratio sample, previously equilibrated with the spike, and 350 ml of 1% m/v sodium tetraethylborate(III) solution are injected simultaneously into the two diVerent lines. The injection volumes, flow rates and conduit lengths are so devised that the two zones precisely overlap each other. The two flows merge in a T-piece where the volatile cadmium species are generated.The gas–liquid mixture formed reaches the gas–liquid separator, where the liquid is drained and the gaseous products are swept by an argon stream through the U-type gas–liquid separator to the torch of the ICP-MS. As described previously,27 this procedure was followed to reduce the amount of the expensive NaBEt4 reagent used during the analysis. Isotope dilution equations Depending on the type of mixing between the sample and the spike, the theoretical equations used to calculate the sample Fig. 2 Optimization of the Einzel 1 and 3 lenses in the HP4500. concentration are diVerent. When the sample and spike are mixed in a batch mode [Fig. 1(a) and 1(c)], the analyte concentration in the sample was calculated using the following equation:9 Cs=Csp Msp Ms Ars Arsp A111sp A114s Rm-Rsp 1-RmRs where Cs is the unknown concentration of cadmium in the sample (s) and Csp the concentration of cadmium in the spike (sp); Ms and Msp are the masses taken from the sample and spike, respectively, Ars and Arsp are the cadmium atomic masses in the sample (natural atomic mass) and spike, respectively, A111sp is the isotope 111 abundance (at.-%) in the spike and A114s is the isotope 114 abundance in the sample (natural abundance), Rm and Rsp are the isotope ratios (114Cd/111Cd) in the mixture and the spike, respectively, and Rs is the isotope Fig. 3 Optimization of the Omega (+) lens in the HP4500. ratio in the sample (111Cd/114Cd). J. Anal. At. Spectrom., 1999, 14, 113–120 115whereas for the Omega (+) lens (Fig. 3) the change is not so factor, K, can be calculated using clear. It was also observed that the best precision was always obtained at the optimum ion lens settings, as expected for K= (Rexp-Rtheo)/Rtheo DM counting statistics. Lens settings which aVected the value of the 111/114 ratio in the HP4500 instrument were Extract 1, were Rexp is the measured isotope ratio, Rtheo is the theoretical Extract 2, Einzel 1 and 3, Einzel 2 and Pole Bias, whereas the isotope ratio and DM is the mass diVerence between the ratio was not clearly aVected by Omega Bias, Omega (+), measured isotopes.In this work, the mass discrimination Omega (-), QP Focus, Ion Deflector and Plate Bias. factor was determined by measuring solutions of natural In order to study the optimum conditions for the precise cadmium and the corrected isotope ratio, Rcorr, was calcumeasurement of Cd isotopic ratios, the eVect of integration lated using.time was also examined using a 25 ng g-1 solution of natural Rcorr=Rexp/(1+K DM) . Cd. The dependence of the relative standard deviation (RSD) (n=5) of the 111Cd5114Cd ratio measurement on the inte- The within-day variation in the mass discrimination factor gration time per unit mass showed that the measured RSD was also examined. As shown in Fig. 5, a significant drift in decreased from 2.5% at a 0.1 s integration time to 0.2% at a the isotope ratio with time was not observed during 2 h of 10 s integration time per unit mass and it becames constant measurement.However, in order to aim at the most accurate for longer integration times, and so an integration time of 10 s results, a mass discrimination correction standard were measper unit mass was selected. ured between every five samples (approximately 30 min). In the on-line ID method, instead of calculating the mass Detector dead-time and mass discrimination errors discrimination factor, the mass bias can be minimized if Rst#Rm (Cst#Cs).22 Therefore, a series of Cd natural stan- It has been shown previously9 that mass discrimination and dards were measured before the measurement of the samples detector dead-time losses could be a serious source of systemand their Cd concentrations were calculated using the standard atic errors in isotope dilution analysis with ICP-MS.that gave the closest isotope ratio to the sample. After an ion generates an electron pulse in the multiplier, there is a finite time during which the system is incapable of Spectral inteferences recording another event; this interval is termed the dead-time.The eVect of the detector dead-time t on the measured count There is always the risk that isotope ratio measurements in rate can be expressed as9 Ireal=Imeas/(1-t Imeas), where Imeas complicated matrices will suVer from isobaric and polyatomic is the experimental count rate and Ireal is the count rate that interferences.12 This potential problem was studied prior to should have been measured if there were no dead-time errors. the blend measurements (by measuring the chosen isotope In order to determine the detector dead-time in our instrument, pairs in unspiked samples).Some workers have reported a we measured a series of diVerent isotope ratios with respect possible MoO+ interference,29 but this was found to be to the isotope 114 at various Cd concentration levels from 10 negligible in our system: the 111Cd5114Cd ratio remained to 100 ng g-1.The results obtained are shown in Fig. 4 as the constant when 1 mg g-1 of Mo was added to 10 ng g-1 of Cd experimental divided by the theoretical isotope ratio versus the (Mo5Cd=15100). concentration of Cd. As can be observed, no change in However, the interference of Sn on 114Cd was important. the ratios with Cd concentration occurred, so a dead-time The presence of 1 mg g-1 of Sn caused an error of about 78% correction was not necesary in this range of concentrations.in the 111Cd5114Cd ratio in a solution of 10 ng g-1 of Cd. This Since concentrations higher than 100 ng g-1 were not meas- Sn interference was corrected using the appropriate correction ured in this work, the dead-time did not need to be taken into equation by measuring the 118Sn isotope (free of isobaric account for the measurement of isotope ratios in the HP4500 interference). instrument. Mass bias is another eVect causing deviations of measured Optimization of spike addition isotopic ratios from the expected ratios.Mass bias derives Taking into account random error propagation, there is a from the diVerential transmission of ions of diVerent mass range of isotope ratios in the mixture which will provide the from the point at which they enter the sampling device until minimum error magnification factor.9 For the isotopic com- they are finally detected by the electron multiplier. As can be position of natural Cd and the 111Cd spike given in Table 2, seen in Fig. 4, there is a clear deviation from the theoretical the measured 1145111 ratio should be between 0.02 and 0.67 ratios (Rexp/Rtheo=1) for all measured Cd isotope ratios. This for an error magnification factor below 1.5. This means that, eVect is most pronounced the larger the mass diVerence if the precision of the measured isotope ratio is typically 0.2%, between the measured isotopes. The mass discrimination Fig. 4 Isotope ratios measured with ICP-MS for four diVerent natural Cd concentrations (only a few isotope ratios of Cd were included to Fig. 5 Drift in 111Cd5114Cd ratio with time. simplify the figure). 116 J. Anal. At. Spectrom., 1999, 14, 113–120and independent of the measured ratio for a large range of precision values obtained are those of the ID methodology. As can be seen, the data obtained for urine are in good values, the expected error in the concentration will be between 0.2 and 0.3% owing to random error propagation.The mini- agreement with the certified value with a precision of 0.5%. However, for the riverine water our data did not agree well mum error corresponded to a 1145111 ratio of 0.12 (error magnification factor of 1.1). with the certified value and the isotope ratio precision is poor. Both negative results can be ascribed to the very low Cd concentrations in the sample (probably requiring pre- Conventional isotope dilution concentration).The digested reference materials were mixed with the spike solution as indicated in Fig. 1 and the 1145111 and 1125111 On-line isotope dilution ratios were measured by ICP-MS. Typically for solid samples 3–4 independent digestions were made and each digestion was The same samples were analysed using the on-line ID method spiked twice to check the precision of both digestion and described previously22 and illustrated in Fig. 1(b). The results spiking procedures.the data obtained for the IAEA H-8 obtained, with reference to a natural Cd standard, are pre- (Horse Kidney sample) are illustrated in Table 3. As can be sented in Table 6 for all materials. When the results presented observed, in all cases the results obtained with the 1145111 in Table 6 are compared with those in Tables 3, 4 and 5, it ratio corresponds to a similar concentration obtained using can be concluded that both procedures provide data of similar the 1125111 ratio.With the exception of sample 2, the quality, the on-line method being simpler, as it does not reproducibility of the spiking procedure seems to be better require the addition of the spike and can be automated using than that of the digestion. On average, the reproducibility of an autosampler.22 the whole procedure was 0.8%, slightly higher than the value expected based on random error propagation only, probably Isotope dilution-FI-vapour generation owing to the irreproducibility of the digestion procedure.The reference value for this material is given as the range Optimization of the Cd signal. Theoretically, vapour generation techniques oVer unique advantages, including elimin- 184–193 mg g-1 Cd. The data obtained for BCR TP-25 Lichens and PACS-1 ation of the need for a nebulizer, enhancement of analyte transport eYciency (approaching 100%), reduction of matrix Marine Sediment are illustrated in Table 4 for the 1145111 ratio only. As can be observed, there is good agreement eVects and the presentation of a homogeneus vapour to the atomizer.30 Moreover, the sampler clogging problem that between the certified and the found values.For PACS-1 the precision seems to be limited also by the sample digestion and occurs when a high salt content sample is analysed is alleviated. It has been shown previously by our group that sodium not by the isotope dilution itself. However, for BCR TP-25 the precision is limited by the measurement of the isotope tetraethylborate(III) is an eYcient derivatization reagent to obtain volatile cadmium species.26 Hence we have applied this ratios because of the low concentration of Cd in the sample.The results obtained for SRM 2670 Freeze-Dried Urine and method for isotope dilution and compared it with the two previously described methods. SLRS-3 Riverine Water are presented in Table 5. For these two samples no sample preparation is necessary and so the First, the parameters aVecting the generation of volatile Table 3 Results obtained for IAEA H-8 Horse Kidney by conventional ID Concentration found/ Concentration found/ Sample 114Cd/111Cd RSD (%) mg g-1 112Cd/111Cd RSD (%) mg g-1 1a 0.1194 0.27 189.2 0.1140 0.31 189.1 1b 0.1207 0.36 189.0 0.1155 0.32 190.0 2a 0.1302 0.34 185.4 0.1233 0.20 186.1 2b 0.1320 0.18 190.0 0.1256 0.22 191.1 3a 0.1205 0.38 187.2 0.1152 0.30 187.7 3b 0.1244 0.21 187.1 0.1184 0.17 187.7 4a 0.1224 0.22 188.6 0.1169 0.33 189.4 4b 0.1222 0.29 189.3 0.1166 0.30 190.4 Average 188.2 188.9 s 1.5 1.7 RSD (%) 0.8 0.9 Table 4 Results obtained for BCR TP-25 Lichens and PACS-1 Marine Sediment by conventional ID BCR TP-25 PACS-1 Concentration/ Concentration/ Sample Ratio RSD (%) mg g-1 Ratio RSD (%) mg g-1 1a 0.1178 1.15 0.860 0.1765 0.57 2.455 1b 0.1186 1.57 0.854 0.1642 0.56 2.347 2a 0.1268 0.94 0.857 0.1615 0.48 2.296 2b 0.1190 0.76 0.852 0.1601 0.38 2.216 3a 0.1207 1.02 0.861 0.1691 0.39 2.214 3b 0.1205 0.58 0.854 0.1724 0.35 2.244 4a 0.1243 1.48 0.866 — — — 4b 0.1326 0.27 0.867 — — — Average 0.859 2.30 s 0.006 0.09 RSD (%) 0.7 3.9 Certified 0.937±0.226 2.38±0.20 J.Anal. At. Spectrom., 1999, 14, 113–120 117Table 5 Results obtained for SRM 2670 Urine and SLRS-3 Riverine Water by conventional ID SRM 2670 SLRS-3 Concentration/ Concentration/ Sample Ratio RSD (%) mg l-1 Ratio RSD (%) mg l-1 1 0.1210 0.32 85.6 0.2437 3.33 0.021 2 0.1206 0.36 86.3 0.2121 3.89 0.024 3 0.1204 0.33 86.3 0.2585 3.81 0.020 4 — — — 0.2516 3.33 0.022 Average 86.1 0.022 s 0.4 0.002 RSD (%) 0.5 7.8 Certified 88±3 0.013±0.002 Table 6 Results obtained for the certified reference materials by on-line ID Concentration Material Sample found IAEA H-8 1a 184.0 Horse Kidneya 1b 183.5 2a 186.0 2b 186.6 3a 185.7 3b 185.5 4a — 4b 185.9 Mean±s 185.3±1.1 RSD (%) 0.6 BCR TP-25 1 0.827 Lichensa 2 0.836 3 0.840 Fig. 6 EVect of (A) HNO3 concentration and (B) NaBEt4 concen- Mean±s 0.834±0.007 tration on the sensitivity of FI-VG of cadmium.RSD (%) 0.8 PACS-1 1a 2.347 Marine Sedimenta 1b 2.298 2a 2.306 2b 2.333 3a 2.336 3b 2.335 Mean±s 2.326±0.019 RSD (%) 0.8 SRM 2670 1 88.2 Freeze-Dried Urineb 2 85.6 3 89.8 4 89.6 Mean±s 88.3±1.9 RSD (%) 2.2 SLRS-3 1 0.029 Riverine Waterb 2 0.035 3 0.039 Mean±s 0.034±0.005 RSD (%) 15 aConcentrations in mg g-1. bConcentrations in mg l-1. Fig. 7 EVect of (A) sample flow rate and (B) Ar flow rate on the sensitivity of FI-VG of cadmium.organometallic Cd species resulting from the merging of an acidified sample stream with sodium tetraethylborate(III) were studied in detail using the merging zones flow injection procedure used before for Bi.27 Maximum peak area at the 111Cd flow rates increased the area of the ion signal but the peak width and the analysis time increased. A final flow rate of intensity was the selected optimization criterion. Two chemical parameters aVect the eYciency of cadmium 3 ml min-1 in each channel was eventually selected.The argon flow rate in the gas–liquid separator was also studied [see vapour generation. First, the eVect of the concentration of HNO3 in the cadmium signal was investigated. The results are Fig. 7(b)] and 1.20 l min-1 was eventually selected. A sample loop of 300 ml was used as a compromise between plotted in Fig. 6(a) and show a critical influence of this acid concentration on the analytical signal with an optimum value signal height and consumption of sample and, particularly, of the expensive borate reagent. Using the merging zones FI of 0.04 M.The eVect of NaBEt4 concentration was also tested and the results are shown in Fig. 6(b). As the NaBEt4 concen- mode it is critical that the two zones of sample and reagent overlap each other precisely. Thus, the borate reagent zone tration increased, the peak area of cadmium signal increased and stabilized at about 1% m/v, so an NaBEt4 concentration was eventually made slightly larger than that of the sample (i.e., 350 ml were selected), to allow slight fluctuations in the of 1% m/v was used in subsequent experiments.Fig. 7 shows the eVect of the flow rates of the sample (A) relative position of the sample and reagent zones at the merging point. and the carrier argon (B) on the peak area. Lower sample 118 J. Anal. At. Spectrom., 1999, 14, 113–120Table 7 Analysis of environmental and biological certified reference materials Concentration found by ID-ICP-MSc Reference Material value Conventional On-line FI-VG SRM-2670 88±3 86± 1 88± 2 87± 4 Freeze-Dried Urinea IAEA H-8 184–193 188±2 185±2 209±13 Horse Kidneyb BCR TP-25 0.937±0.226 0.859±0.006 0.834±0.007 — Lichensb SLRS-3 0.013±0.002 0.022±0.002 0.034±0.005 — Riverine Watera PACS-1 2.38±0.09 2.30±0.09 2.33±0.02 2.54±0.61 Marine Sedimentb aConcentrations in mg l-1.bConcentrations in mg g-1. cMean±s. Interference studies. The eVect of the presence of foreign dilution and the mass bias error can be easily minimized if Rst=Rm.Further, the measurement of isotopic ratios of a elements on the vapour generation of cadmium in the proposed procedure was investigated. Potentially interfering elements in natural Cd standard between the samples allows one to perform ID analysis (referring the results to the natural Cd the previous ethylide generation experiments26,31,32 were tested; Cu, Co and Ni, were found not to aVect the generation of standard) regardless of the concentration of the spike.Unfortunately, this method requires a larger spike cadmium ethylide to a great extent, even at relatively high levels (recoveries of around 105–115% were observed for consumption. Finally, ID-FI-vapour generation using NaBEt4 did not 10 ng g-1 of cadmium in presence of 5 mg g-1 of these interferents). provide results as good as expected. The sensitivity did not increase and the precision was worse than those of the other The eVect of a high salt content matrix (e.g., as in seawater) on the recovery of cadmium was also tested.A 100% two approaches. However, the approach could be useful for determinations of cadmium in high salt content samples with recovery was found for spiked undiluted sea-water with 10 ng g-1 of cadmium. low consumption of sample and spike. Analysis of reference materials. The sensitivity of this Acknowledgements FI–vapour generation of cadmium method did not increase in The loan of the HP4500 ICP-MS instrument by Hewlett- comparison with conventional nebulization of cadmium in an Packard and Project DG-94-PB-1331 from DGICYT FI system.The within-run precision was determined using five (Madrid) are gratefully acknowledged. injections of a solution containing 10 ng g-1 of cadmium. The RSDs of the peak areas and isotope ratio measurements were 7.8 and 2.4%, respectively. Unfortunately, because of the poor References isotope ratio precision, the expected relative error in the case 1 Handbook on Toxicity of Inorganic Compounds, ed.H. G. Seiler of isotope dilution will be 10 times that obtained by convenand H. Sigel, Marcel Dekker, New York, 1988. tional nebulization. 2 H. J. Yang, K. S. Huang, S. J. Jiang, C. C. Wu and C. H. Chou, The final results obtained by merging zones FI vapour Anal. Chim. Acta, 1993, 282, 437. generation isotope dilution for three of the reference materials 3 L. Ebdon, A.S. Fisher, P. J. Worsfold, H. Crews and M. Baxter, are presented in Table 7, compared with the values obtained J. Anal. At. Spectrom., 1993, 8, 691. 4 D. E. Nixon and T. P. Moyer, Spectrochim. Acta, Part B, 1996, by the other two ID methodologies and the reference values. 51, 13. As can be observed, both the precision and accuracy are worse 5 J. M. Cook, J. J. Robinson, S. R. N. Chenery and D. L. Miles, with this FIA method, with the exception of the urine sample. Analyst, 1997, 12, 1207.Further work in order to increase the sensitivity of the Cd 6 H. Kumagai, M. Yamanaka, T. Sakai, T. Yokoyama, T. M. vapour generation system and to decrease the width of the Suzuki and T. Suzuki, J. Anal. At. Spectrom., 1998, 13, 579. peaks obtained will be neccesary in future studies using 7 K. G. Heumann, in Inorganic Mass Spectrometry, ed. F. Adams, R. Gijbels and R. J. Grieken, Wiley, New York, 1988, p. 301. alternative FIA approaches. 8 J. R. Moody and M.S. Epstein, Spectrochim. Acta, Part B, 1991, 46, 1571. Conclusions 9 J. I. Garcý�a-Alonso, Anal. Chim. Acta, 1995, 312, 57. 10 IPapadakis, P. D. P. Taylor and P. De Bie`vre, Anal. Chim. Acta, Ultratrace cadmium determination in biological and environ- 1997, 346, 17. mental samples requires the best possible precision and accu- 11 T. J. Hwang and S. J. Jiang, J. Anal. At. Spectrom., 1997, 12, 579. 12 J. W. McLaren, D. Beauchemin and S. S. Berman, Anal. Chem., racy. In this context, ID-ICP-MS can provide highly precise 1987, 59, 610.and accurate results at very low concentration levels of cad- 13 D. Beauchemin, J. W. McLaren, A. P. Mykytiuk and S. S. mium. Moreover, systematic errors (mass bias eVects and Berman, Anal. Chem., 1987, 59, 778. spectroscopic interferences) can be easily corrected using the 14 P. L. Lu, K. S. Huang and S. J. Jiang, Anal. Chim. Acta, 1993, software of the instrument. 284, 181. Validation experiments have shown that the two new 15 J.Wu and E.A. Boyle, Anal. Chem., 1997, 69, 2464. 16 T. Katoh, M. Akiyama, H. Ohtsuka, S. Nakamura, K. Haraguchi approaches, compared here with the conventional ID analysis, and K. Akatsuka, J. Anal. At. Spectrom., 1996, 11, 69. can be used for the determination of ultratrace concentrations 17 K. G. Heumann, Anal. Chim. Acta, 1993, 283, 230. of cadmium in real samples of biological and environmental 18 T. J. Hwang and S. J. Jiang, J. Anal. At. Spectrom., 1996, 11, 353. interest. 19 C. C. Chan and S. J. Jiang, J. Anal. At. Spectrom., 1997, 12, 75. On-line ID has turned out to be more convenient for the 20 K. H. Lee, S. H. Liu and S. J. Jiang, Analyst, 1998, 123, 1557. determination of cadmium, because it provides similar results 21 C. J. Park and J. K. Suh, J. Anal. At. Spectrom., 1997, 12, 573. 22 J. M. Marchante-Gayo�n, J. I. Garcý�a Alonso and A. Sanz-Medel, to conventional ID, but avoids the tedious step of oV-line J. Anal. At. Spectrom., 1999, 14, 113–120 119in Plasma Source Mass Spectrometry: Developments and 28 P. J. De Bie`vre and P. D. P. Taylor, Int. J. Mass Spectrom. Ion Applications, ed. G. Holland and S. D. Tanner, Royal Society of Processes, 1993, 123, 149. Chemistry, Cambridge, 1997, pp. 85–94. 29 C. W. McLeod, A. R. Date and Y. Y. Cheung, Spectrochim. Acta, 23 A. La�sztity, M. Viczia�n, X. Wang and R. M. Barnes, J. Anal. At. Part B, 1986, 41, 169. Spectrom., 1989, 4, 761. 30 J. Dedina and D. L. Tsalev, Hydride Generation Atomic Absorption 24 M. Viczia�n, A. La�sztity, X. Wang and R. M. Barnes, J. Anal. At. Spectrometry,Wiley, New York, 1995. Spectrom., 1990, 5, 125. 31 A. D’Ulivo and Y. Chen, J. Anal. At. Spectrom., 1989, 4, 319. 25 H. Klinkenberg, T. Beeren, W. Van Borm, F. van der Linden and 32 L. Ebdon, P. Goodall, S. J. Hill, P. B. Stockwell and K. C. M. Raets, Spectrochim. Acta, Part B, 1993, 48, 649. Thompson, J. Anal. At. Spectrom., 1993, 8, 723. 26 M. C. Valde�s-Hevia y Temprano, M. R. Ferna�ndez de la Campa and A. Sanz-Medel, J. Anal. At. Spectrom., 1994, 9, 231. 27 J. P. Valles Mota, M. R. Ferna�ndez de la Campa and A. Sanz- Medel, J. Anal. At. Spectrom., 1998, 13, 431. Paper 8/07361I 120 J. Anal. At. Spectrom., 1999,
ISSN:0267-9477
DOI:10.1039/a807361i
出版商:RSC
年代:1999
数据来源: RSC
|
4. |
The use of ETV-ICP-MS for the determination of selenium in serum |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 121-126
Justine Turner,
Preview
|
|
摘要:
The use of ETV-ICP-MS for the determination of selenium in serum†‡ Justine Turner,*a Steve J. Hill,b E. Hywel Evansb and Ben Fairmana aLGC (Teddington) Ltd., Queens Road, Teddington, Middlesex UK, TW11 OLY bUniversity of Plymouth, Department of Environmental Sciences, Drake Circus, Plymouth, Devon UK, PL4 8AA Received 24th August 1998, Accepted 26th October 1998 The development of a novel procedure for the accurate determination of selenium in serum using electrothermal vaporisation inductively coupled plasma mass spectrometry (ETV-ICP-MS) is described.The proposed method eliminates the need for a lengthy sample digestion procedure (a requirement with many methods for the analysis of biological samples), utilising a simple 1+19 dilution of the serum with 1% nitric acid.Many of the interferences normally associated with the determination of selenium by ICP-MS are successfully eliminated with careful optimisation of the ETV temperature program and modifier system.Analytical characteristics for 74Se, 77Se, 78Se and 82Se are reported, including detection limits (3s blank) of approximately 0.1 ng g-1 for 77Se and 82Se. Short- and longterm reproducibility data between 4.7 and 4.9% and 3.2 and 3.8% (RSD) for 77Se and 82Se, respectively, are shown. The accuracy of the method, which included Te as an internal standard, was demonstrated with the analysis of three internal quality control samples and the certified reference material NIST SRM1598 (bovine serum).Results within 10% of the target value were achieved for three of the four isotopes studied, with slightly worse results for 78Se owing to the large interference from argon adduct ions on this isotope. Preliminary work involving the addition of nitrogen to the argon aerosol carrier gas was successful in reducing the ArAr interference at m/z 78. levels of components such as sodium and organic compounds, Introduction for instance proteins, can cause signal suppression. Several Selenium is an essential trace element whose involvement in workers4–10 have used hydride generation techniques to overhuman health and well being has become evident in recent come some of the problems.Greater sensitivity is attainable years. Since the 1930s the perception of selenium has gone owing to an improved sample delivery rate, and reduction of through a number of changes, namely from being considered spectroscopic and non-spectroscopic interferences is achieved a toxic element, to a carcinogen, to an essential element and as a result of analyte removal from the matrix.However then in the 1960s and 1970s to being considered an anti- lengthy sample preparation procedures are generally required, carcinogen,1 thus illustrating the marginal diVerences between to breakdown organic selenium in the matrix and for the therapeutic and toxic eVect. Both an excessive and insuYcient conversion of SeVI into SeIV prior to generation of the hydride.intake of selenium can have serious health implications. Another option for the reduction of interferences involves the Selenium toxicity, selenosis, can be fatal, with symptoms such addition of an organic solvent to the diluent matrix when as hair and nail loss, tooth decay, skin lesions and, in severe using conventional nebulisation ICP-MS,11–13 although probcases, abnormalities of the nervous system. Deficiencies have lems of nebuliser blockage associated with the viscous serum been linked with coronary heart disease, acute myocardial matrix still need to be addressed.Electrothermal vaporisation infarction, cirrhosis of the liver2 and cancer of several major (ETV) sample introduction is an alternative approach, with organs including the large intestine, breast, ovary and lungs. the potential to eliminate some of the polyatomic interferences Considering the health implications detailed and the narrow already mentioned.Other advantages over conventional and divide between deficiency and toxicity, it is essential that good, hydride generation procedures include small sample sizes accurate and reliable methods of analysis are available. (5–50 ml ), minimal sample pre-treatment, improved sensitivity ICP-MS is widely used in many routine analytical labora- and low absolute detection limits. tories. Advantages over rival techniques include low detection This paper describes the development of an ETV-ICP-MS limits and speed of analysis. However not all determinations procedure for the determination of Se in serum.Optimisation are straightforward, in particular, the determination of of the ETV temperature program including the successful selenium in serum is complicated by several factors. The elimination of several interferences, together with evaluation sensitivity achieved with conventional ICP-MS is generally of the procedure using a certified reference material, is poor: owing to the high first ionisation energy of selenium, described in detail.Preliminary work using an N2–Ar mixed only 30% ionisation is achieved with an argon plasma;3 plasma gas to negate ArAr polyatomic interferences is also spectroscopic interferences caused by the formation of argon discussed. polyatomic species, i.e., 40Ar37Cl on 77Se and 40Ar38Ar on 78Se, lead to high background levels, poor detection limits and Experimental ultimately biased analytical results; and the presence of high Instrumentation An ELAN 5000A ICP-MS instrument coupled to an HGA †Presented at the Ninth Biennial National Atomic Spectroscopy 600MS ETV unit with an AS-60 autosampler attachment Symposium (BNASS), Bath, UK, July 8–10, 1998.‡© Copyright LGC (Teddington) Ltd. 1998. (Perkin-Elmer, Beaconsfield, UK) was used. A 140 cm long J. Anal. At. Spectrom., 1999, 14, 121–126 121Table 1 ICP-MS operating conditions (Alfa, Johnson Matthey, Royston, UK) were used.Working standards were prepared daily by dilution in 1% m/m HNO3, ICP— ultrapure Ultrex II grade acid [JT Baker (UK), Milton Power 1150 W Keynes, Buckinghamshire, UK]. Palladium(II) nitrate (Sigma, Outer plasma gas 15.0 l min-1 Poole, Dorset, UK) was used to prepare the chemical modi- Intermediate gas 0.80 l min-1 Aerosol carrier gas 0.95 l min-1 fier solution. Cones Pt Lenses P 48 Preparation of samples and solutions B 43 S 45 Chemical modifier. A 500 mg g-1 solution of Pd(NO3)2 E 25 containing 5 ng g-1 Te (internal standard) was prepared in the following way: 0.05 g of Pd(NO3)2 was dissolved in 50 g Parameter file— of 10% HNO3; 0.5 g of Te standard (1.0 mg g-1) was then Dwell time 15 ms added and the solution diluted to 100 g with 1% HNO3.Sweeps/reading 1 Readings/replicate 60 Points across peak 1 Samples. All serum samples were diluted (1+19) with 1% Resolution Normal HNO3 (Baker). Mass 74Se 77Se 78Se Results and discussion 82Se Chemical modifiers The use of chemical modifiers with electrothermal techniques is well established.In ICP-MS an enhancement in signal on addition of a chemical modifier is attributed to a more eYcient transport of the vaporised analyte to the plasma.14 Chemical modification is also important to avoid losses of volatile analytes during the ashing stage with the formation of more stable analyte species.15 In this study numerous chemical modifiers were examined including: Pd(NO3)2, Pd(NO3 )2+ Ni(NO3)2, Pd(NO3 )2+Mg(NO3)2, ascorbic acid and Pd(NO3)2+Mg(NO3)2+ascorbic acid.The best results were obtained with a Pd(NO3)2 only modifier. The optimum concentration was established by examination of the changes in Se signal of a 10 ng g-1 standard in 1% HNO3 with increasing Pd(NO3)2 concentration. Findings indicated that 10 ml of a 100 mg g-1 Pd(NO3)2 solution were consistent with a maximum signal. However, further work with a serum sample showed evidence of selenium losses during the pyrolysis stage at temperatures above 1000 °C.Increasing the modifier Fig. 1 Pyrolysis/vaporisation curves for 74Se, 77Se, 78Se and 82Se; 10 ml injection of a 10 ng g-1 standard with 10 ml of 500 mg g-1 Pd(NO3)2 concentration to 500 mg g-1 successfully overcame this modifier. problem with little eVect on overall sensitivity. Optimisation of ETV temperature program piece of PTFE tubing (0.6 cm id) was used to connect the furnace to the ICP-MS. Operating conditions for the ICP-MS L’vov platform.The work discussed in this study was instrument and the ETV temperature program are given in performed using L’vov platformpyrolytic graphite coated graph- Tables 1 and 2, respectively. Optimisation of the ICP-MS ite tubes. The advantages of this type of tube have been docu- instrument (i.e., lens settings, resolution, oxide and doubly mented by several workers16,17 who have described the platform charged ion formation) was performed using conventional furnace tube at a stabilised ‘steady state’ temperature.A com- nebulisation prior to coupling the instrument to the ETV unit. parison was made between this and a non-platform tube with the determination of a 10 ng g-1 Se standard at increasing Reagents pyrolysis temperatures. Data showed that despite an eventual decrease in signal with an increase in pyrolysis temperature, a All solutions were prepared using high purity deionised water (18MV, Elga, High Wycombe, Buckinghamshire, UK).Stock more consistent signal was obtained with the L’vov tube than with a non-platform graphite tube. This supports the work of solutions (1000 mg ml-1) of Se and Te (internal standard) Table 2 ETV temperature program Sample volume 10 ml Injection speed 60% Modifier volume 10 ml Read delay 2.5 s Gas/l min-1 Step Temperature/°C Ramp/s Hold/s Internal External Read Dry 1 110 10 15 0.3 Dry 2 120 10 45 0.3 Pyrolysis 1100 10 45 0.3 Vaporisation 2600 0.5 1 0.3 0.95 Yes Clean 2700 0.0 1 0.95 Cool 20 15 1 0.95 122 J.Anal. At. Spectrom., 1999, 14, 121–126removal ) and vaporisation (dissociation of atoms) processes. Parameters at each of these stages, such as ramp rate, temperature and hold time, were evaluated to establish the optimum conditions. Pyrolysis and vaporisation temperatures were optimised by the repeated analysis of a 10 ng g-1 Se standard at increasing temperature settings. Temperature curves constructed with the data from these experiments can be seen in Fig. 1. For the optimisation of the pyrolysis temperature the vaporisation temperature was set at 2600 °C, and for the vaporisation temperature experiment, the pyrolysis temperature was set at 1100 °C. The pyrolysis curve shows a stable signal between 500 and 1300 °C, and a decrease in signal at temperatures above this, suggesting that a pyrolysis temperature within the range mentioned would be suitable. Owing to the fact that the graphite tube degrades with the number of firings, it was decided that a pyrolysis temperature midway in the range rather than at the higher end would be chosen, to Fig. 2 Pyrolysis curves for 77Se, 82Se and 128Te; 10 ml injection of a minimise the detrimental eVect on the lifetime of the tube. A serum sample (with an approximate Se concentration of 5 ng g-1); temperature of 800 °C was selected and a repeat experiment 10 ml of a 500 mg g-1 Pd(NO3)2 modifier were used. using a serum sample performed.The response of the 77Se, 82Se and 128Te (internal standard) signals with an increase in temperature in the presence of the serum matrix is shown in Slavin et al.,15 who demonstrated the diVerences between analyte Fig. 2. Suppression of the Te signal, and to a lesser extent the atomisation from the furnace wall and the platform. Se signal, can be seen at lower temperatures. Serum contains high concentrations of components such as sodium, chlorine Pyrolysis and vaporisation.The main stages of any electrothermal vaporisation program are the pyrolysis (matrix and bromine. Examination of these analytes alongside Se in Fig. 3 The eVect of pyrolysis temperature with respect to interference elimination; 10 ml injection of a serum sample in 1% HNO3, approximate concentration 5 ng g-1. (a) 77Se and 23Na signal profiles, pyrolysis temperature 800 °C; (b) 77Se and 23Na signal profiles, pyrolysis temperature 1200 °C; (c) 77Se and 35Cl signal profiles, pyrolysis temperature 800 °C; (d) 77Se and 35Cl signal profiles, pyrolysis temperature 1200 °C; (e) 82Se and 79Br signal profiles, pyrolysis temperature 800 °C; and (f ) 82Se and 79Br signal profiles, pyrolysis temperature 1200 °C.J. Anal. At. Spectrom., 1999, 14, 121–126 123are shown in Fig. 3(a). The detector was desensitised at m/z 23 using the Omnirange option in the ELAN software. This enabled the Na signal to be plotted on the same axis as the Se in order to gain a direct comparison between the two signals.As can be seen, the Na signal coincides directly with the Se signal, resulting in large amounts of Na ions in the ETV transfer line and plasma at the same time as the analyte of interest. This could potentially aVect the transport and ionisation eYciency of Se, thus causing a suppression of the signal. An increase in the pyrolysis temperature to 1200 °C successfully separated the two signals, with the Na burning oV at an earlier point [see Fig. 3(b)]. Similar responses were observed with Cl and Br signals. The problems presented by Cl on 77Se are illustrated in Fig. 3(c). The formation of 40Ar37Cl ions in the plasma would enhance the signal at m/z 77, coinciding with the 77Se isotope, leading to high background signals, poor detection limits and biased analytical results. From the diagrams it can be seen that at 800 °C the Cl signal profile overlaps with the Se signal. An elevated pyrolysis temperature of 1200 °C separated the Cl from the Se resulting in the interference free determination of 77Se [Fig. 3(d)]. Bovine serum (used to prepare the IQC samples and NIST SRM 1598 discussed later) contains high levels of bromine. Fig. 4 The eVect of vaporisation temperature on 77Se signal profile; The combination of 81Br with hydrogen produces HBr with 10 ml injection of a serum sample in 1% HNO3, approximate concenan m/z of 82, coinciding with the 82Se isotope. Again at 800 °C tration 5 ng g-1.(a) Vaporisation temperature of 2300 °C; and (b) the Br signal overlaps with the Se signal [Fig. 3(e)], but is vaporisation temperature of 2600 °C. successfully removed at the higher temperature of 1200 °C [Fig. 3(f )]. The degree of interference from HBr is further reduced by the elimination of water vapour. With conventional nebulisation the sample is introduced as an aqueous solution, resulting in a considerable amount of hydrogen and oxygen ions in the plasma. With the ETV, the sample is introduced as a gas, hence the level of hydrogen and oxygen is greatly reduced.The advantage of hydrogen and oxygen reduction with ETV sample introduction has also been discussed by Marshall and Franks.19 As shown in Fig. 1, the Se signal remained fairly constant throughout the vaporisation temperature range examined. Using 2300 °C as the vaporisation temperature, Fig. 4(a) shows the signal profile for a serum sample diluted in 1% HNO3. A large analyte peak is observed at 60 s, but a second much smaller peak is seen slightly later at 70 s.During the temperature program (Table 2) the graphite tube is rapidly heated to Fig. 5 Typical signal profiles for a 10 ml injection with 10 ml of 2700 °C after the vaporisation stage to remove any residual 500 mg g-1 Pd(NO3)2 modifier. (a) 5 ng g-1 Se standard in 1% HNO3; matrix components. The second peak is probably Se burning and (b) diluted serum sample, approximate concentration 5 ng g-1.oV as the furnace temperature increases, suggesting that the temperature of 2300 °C is insuYcient for complete vapor- Table 3 Analytical performance characteristics isation. This is verified in Fig. 4(b) where the vaporisation temperature has been increased to 2600 °C. Only one peak is Parameter 74Se 77Se 78Se 82Se now observed indicating that all of the Se has been vaporised and transported to the ICP-MS. Detection limit/ng g-1 0.85 0.14 0.58 0.13 Absolute detection 8.5 1.4 5.8 1.3 limit/pg Instrument parameters Short-term stability ±15 ±4.9 ±4.6 ±3.2 (n=10) (%) Operating parameters including plasma power, aerosol carrier Long-term stability ±13 ±4.7 ±5.7 ±3.8 gas flow rate and lens settings were optimised.An increase in (n=10) (%) signal intensity for all isotopes was observed with an increase Linearity/ng g-1 1–100 0–100 0–100 0–100 in plasma power, up to a maximum of 1150 W. Above this the signal begins to decrease. All further work was performed at 1150 W, a higher setting than that favoured by other the proposed system gave clear evidence of their role as potential interferents.workers11 when using conventional nebulisation. This can improve the ionisation of Se, and hence the sensitivity, but on Sodium is an easily ionised element (EIE) which can aVect the mass transport eYciency of analytes, the amount of analyte the downside also increases the ionisation of polyatomic species. However with the proposed system the interferences that reaches the plasma and ultimately the signal intensity.O’Hanlon et al.18 investigated the eVect of several EIEs with on 77Se and 82Se have already been eliminated during the ETV process, and therefore this improvement in ionisation produces a plasma emission system, demonstrating a reduction in the transport eYciency of Mn in the presence of Na. The signal a net increase in sensitivity. The eVect of the argon carrier gas flow rate was also investigated.The Se signal increased steadily profiles of 77Se and 23Na at a pyrolysis temperature of 800 °C 124 J. Anal. At. Spectrom., 1999, 14, 121–126Table 4 Accuracy data from the analysis of internal quality control samples and NIST SRM 1598 (bovine serum). Values are expressed as the mean and standard deviations of three measurements Concentration/ng g-1 IQC 1 IQC 2 IQC 3 NIST SRM 1598 Isotope (54.5 ng g-1) (103 ng g-1) (148 ng g-1) (42.4±3.5 ng g-1) 74Se 48.9±6.9 97.9±7.2 146±5.8 37.9±5.2 77Se 51.1±2.9 96.7±4.0 142±7.0 41.4±0.75 78Se 64.1±9.3 109±7.8 146±8.0 55.6±0.72 82Se 50.7±2.2 94.3±4.4 138±4.3 40.8±0.45 with an increase in gas flow rate from 0.85 l min-1, reaching investigate the eVect of nitrogen in the proposed system.Nitrogen was connected to the ICP-MS via the oxygen inlet a maximum at 0.95 l min-1, and decreased at flow rates above this. It should be noted that the internal flow rate of Ar in tube and the nitrogen level regulated using the oxygen mass flow controller.A T-piece fitted in the PTFE transfer tube the furnace is 0.30 l min-1. This combined with the ICP-MS carrier gas optimum flow rate of 0.95 l min-1 leads to a total allowed the nitrogen to mix with the argon before arriving at the plasma. A decrease in the blank level of 78Se was observed carrier gas flow rate of 1.25 l min-1, a similar optimum rate to that reported previously.20 with an increase in nitrogen content. However the signal intensity of a 10 ng g-1 Se standard also decreased in a similar manner.The rate of reduction of the blank and standard Analytical performance signals is illustrated in Fig. 6, along with the variation in the Typical injection profiles for a 5 ng g-1 Se standard and a signal to blank ratio as the nitrogen content increases. From serum sample with an approximate concentration of 5 ng g-1 the graph it can be seen that aerosol carrier gas containing are shown in Fig. 5. The sensitivity of the system corresponds 4% nitrogen gave rise to the largest decrease in the blank to between 200 and 6000 counts per ng g-1 of Se depending level.This study has shown the ability of nitrogen to reduce on the isotopic abundance. the interference at m/z 78, but further work is required to improve the sensitivity of the system. At present the nitrogen Linearity. The system was found to be linear from 0 to is introduced to the aerosol carrier gas after vaporisation of 100 ng g-1 for the 77, 78 and 82 isotopes but only linear from the analyte. An alternative procedure may be to mix the argon 1 to 100 ng g-1 for 74Se. This non-linearity below 1 ng g-1 for and nitrogen prior to the vaporisation stage and use it to carry 74Se may be attributed to its low isotopic abundance (0.90%) the analyte from the ETV into the plasma.and hence the lack of sensitivity. Conclusion Reproducibility. Short-term stability data for ten consecutive analyses (triple firings) of a diluted serum sample, followed by The procedure described enables the accurate determination ten determinations over a 4 h period to give the long-term of Se in serum to be performed with minimal sample prestability of the system, are detailed in Table 3.Data were treatment. The main issue addressed in this study was one of obtained with the intensity ratioed to the Te internal standard. interference elimination. The proposed method has successfully achieved this objective, allowing the interference free deter- Detection limits.Typical limits of detection (calculated as mination of two of the isotopes of selenium—77 and 82. 3s based on ten determinations of 1% HNO3 as the blank) Further reduction of interference on other selenium isotopes are also found in Table 3. The poor detection limits of 74Se has also been shown to be possible using mixed gas plasmas. and 78Se are due to the low abundance and poor sensitivity of 74Se, and the substantial interference from argon polyatomics The work carried out in this paper was supported by the on 78Se.Further work to improve the detection limit of 78Se Department of Trade and Industry as part of the National will continue, with the addition of nitrogen to the argon gas. Measurement System Valid Analytical Measurement Programme. The authors would also like to thank Dr. Trevor Accuracy. To check on the accuracy of the method a number Delves of Southampton University who kindly donated the of internal quality control (IQC) sera (prepared by the IQC serum samples used in this work.addition of Se standards to bovine serum, donated by H. T. Delves, Southampton University) and NIST SRM 1598 (bovine serum) were analysed. The results can be found in Table 4. A linear calibration was performed utilising the blank correction facility in the ELAN software. Excellent agreement between the results obtained and the target values for 74Se, 77Se and 82Se are shown.High results were obtained with 78Se, but again this is attributed to the large interference from argon adduct ions at m/z 78. The RSDs calculated from triplicate analyses of each sample were between 3.1 and 5.7% for 77Se and 82Se and 4.0 and 14.0% for 74Se and 78Se in the IQC samples, and between 1.1 and 1.8% for NIST SRM 1598, with the exception of 74Se which gave an RSD of 14.0%. Nitrogen addition The introduction of nitrogen to argon plasmas and its ability to reduce polyatomic ion formations have been documented Fig. 6 The eVect of nitrogen addition to the argon aerosol carrier gas on the 78Se signal. by several workers.13,21,22 An experiment was performed to J. Anal. At. Spectrom., 1999, 14, 121–126 12511 H. T. Delves and C. E. Sieniawska, J. Anal. At. Spectrom., 1997, References 12, 387. 12 J. Goossens, F. Vanhaecke, L. Moens and R. Dams, Anal. Chim. 1 Handbook on Metals in Clinical and Analytical Chemistry, ed. Acta, 1993, 280, 137.H. G. Seiler, A. Sigel and H. Sigel, Marcel Dekker, New York, 13 E. H. Evans and L. Ebdon, J. Anal. At. Spectrom., 1989, 4, 299. 1994. 14 D. C. Gregoire, S. Al-Maawali and C. L. Chakrabarti, 2 J. Aaseth, J. Alexander, Y. Thomassen, J. P. BloomhoV and Spectrochim. Acta, Part B, 1992, 47, 1123. S. Skerde, Clin. Biochem., 1982, 15, 281. 15 W. Slavin, D. C. Manning and G. R. Carnrick, At. Spectrosc., 3 R. S. Houk, Anal. Chem., 1986, 58, 97A. 1981, 2, 137. 4 M. P. Rayman, F. R. Abou-Shakra and N. I. Ward. J. Anal. At. 16 G. F. Kirkbright, S. Hsiao-Chuan and R. D. Snook, At. Spectrom., 1996, 11, 61. Spectrosc., 1980, 1, 85. 5 M. Haldimann, T. Y. Venner and B. Zimmerli, J. Trace Elem. 17 J. M. Ottaway, At. Spectrosc., 1982, 3, 89. Med. Biol., 1996, 10, 31. 18 K. O¡�Hanlon, L. Ebdon and M. Foulkes, J. Anal. At. Spectrom., 6 M. A. Quijano, A. M. Gutie¢¥rrez, M. Pe¢¥rez Conde and C. Ca¢¥mara, 1997, 12, 329. J. Anal. At. Spectrom., 1995, 10, 871. 19 J. Marshall and J. Franks, At. Spectrosc., 1990, 11, 177. 7 B. T. G. Ting, C. S. Mooers and Janghorbani, Analyst, 1989, 20 B. Fairman and T. Catterick, J. Anal. At. Spectrom., 1997, 12, 863. 114, 667. 21 F. Laborda, M. J. Baxter, H. M. Crews and J. Dennis, J. Anal. At. 8 O. Mestek, M. Sucha¢¥nek, Z. Vodic¢§kova, B. Zemanova¢¥ and Spectrom., 1994, 9, 727. T. Z©¥¢¥ma, J. Anal. At. Spectrom., 1997, 12, 85. 22 T. van der Velde-Koerts and J. L. M. de Boer, J. Anal. At. 9 J. Bowman, B. Fairman and T. Catterick, J. Anal. At. Spectrom., Spectrom., 1994, 9, 1093. 1997, 12, 313. 10 R. M. Olivas, C. R. Que¢¥tel and O. F. X. Donard, J. Anal. At. Spectrom., 1995, 10, 865. Paper 8/06638H 126 J. Anal. At. Spectrom., 1999, 14, 121.126
ISSN:0267-9477
DOI:10.1039/a806638h
出版商:RSC
年代:1999
数据来源: RSC
|
5. |
Determination of mercury in hair by coupled CVAA-ICP-MS |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 127-129
Robert Knight,
Preview
|
|
摘要:
Determination of mercury in hair by coupled CVAA-ICP-MS† Robert Knight,a* Stephen J. Haswell,a Stephen W. Lindowb and Joy Battyb aChemistry Department, University of Hull, Hull HU6 7RX, UK bDepartment of Obstetrics and Gynaecology, Hull Maternity Hospital, Hedon Road, Hull HU9 5LX, UK Received 2nd July 1998, Accepted 20th October 1998 This paper describes the coupling of a cold vapour atomic absorption (CVAA) instrument with an ICP-MS to obtain mercury concentration information and stable isotopic ratio data.The sample preparation procedure, based on microwave digestion, describes the decomposition of small masses of hair using 7 ml digestion vessels, which will enable the methodology to be used subsequently with neonatal hair samples. The results based on a CRM hair sample and spiked real samples indicated that the digestion method with direct CVAA analysis gave recoveries of around 100% with a corresponding RSD of 3.5%. The interfacing of the CVAA instrument to the ICP-MS was relatively simple and enabled isotopic data to be obtained in tandem with CVAA quantitative data. Some ICP-MS mercury memory problems were observed with the coupled CVAA-ICP-MS but these did not aVect either the CVAA determination or isotopic ratio estimations. Introduction Experimental Reagents This paper forms part of a larger study, which involves the monitoring of trace elements in maternal and neonatal hair Nitric acid was SpA grade from Romil Chemicals (Shepshed, samples, to establish causal links with a range of clinical Leicestershire, UK) and the hydrochloric acid was analytical- syndromes.Whilst mercury uptake, from dietary sources and reagent grade obtained from Fisons (Loughborough, UK). dental amalgams, has been previously studied, owing to the The 1000 ppm mercury standard solution was Aristar grade potential health aspects of such exposure,1 information on a (BDH Merck, Poole, Dorset, UK) and was used to generate wider range of elements would clearly be advantageous for the required standards.The tin chloride, AnalaR grade (BDH this present study. Earlier work specifically on mercury indi- Merck), was used as a 10% (m/v) solution in 10% (v/v) cated that determinations by conventional ICP-MS on referhydrochloric acid. Magnesium perchlorate, AnalaR grade ence hair can lead to recoveries as low as 56% of their certified (BDH Merck), was used as a drying agent for the optical value.2 Thus the direct determination of mercury by ICP-MS cells.The diluent water was obtained from an Elga (High as part of a simultaneous analysis of a hair digest, based on Wycombe, Buckinghamshire, UK) UHQ PS system. pneumatic nebulisation, is generally considered to be unre- The CRM human hair, GBW 07601 was obtained from liable. However, the addition of a chemical modifier such as Glenn Spectra (Stanmore, Middlesex, UK). All glassware was hydrobromic acid has been demonstrated to reduce surface grade A, acid washed and soaked in 2% nitric acid. All adsorption of mercury inside the sample introduction system solutions were immediately transferred into polypropylene so eVectively reducing analyte loss.3 Whilst the addition of a vessels after making up to volume.chemical modifier has distinct advantages in determining a particular element, in this case mercury, it can inadvertently Instrumentation lead to elemental contamination and unexpected matrix inter- Sample digestion equipment.A CEM Model MDS81D ferences when multielemental determinations are attempted. microwave oven with a separate pressure control line attached In this work an alternative approach, which alleviates the need to one vessel was supplied by CEM Microwave Technology to add chemical modifiers to the sample, has been used [cold (Buckingham, UK). Samples were digested in 7 ml microves- vapour atomic absorption (CVAA)], but at the cost of losing sels made from Teflon, placed inside Teflon PFA advanced stable isotopic data.The ability to obtain isotopic ratio analysis composite vessels (ACV). can be of great value in clinical studies,4 for example it would enable the successful sourcing of isotopically labelled dental ICP-MS. A VG Elemental (Winsford, Cheshire, UK) amalgam to be carried out on non-invasive samples such as PQ2+ ICP-MS was used. The conventional nebulisation hair. In this work CVAA has been coupled with ICP-MS to arrangement included a Cetac MCN 100 microconcentric create an analytical approach capable of generating both nebuliser (Cetac Technologies, Crewe, Cheshire, UK) with an quantitative and stable isotopic ratio analysis for mercury.In uptake of 35 ml min-1 and a quartz Scott type double pass addition an important aspect of the methodology developed spray chamber chilled to 4 °C. The instrument was operated will be the capability to deal with small sample masses of under the following conditions: aerosol carrier flow neonatal baby hair in order to obtain simultaneous 0.88 l min-1, coolant gas flow 13.0 l min-1, intermediate flow multielemental data for small sample volumes. 0.9 l min-1, forward power 1350 W, reflected power <3W. Mercury was determined in triplicate at m/z 200 and 202 in peak jumping mode, 3 points per peak, 10.24 ms per point †Presented at the Ninth Biennial National Atomic Spectroscopy Symposium (BNASS), Bath, UK, July 8–10, 1998. for 30 s.J. Anal. At. Spectrom., 1999, 14, 127–129 127CVAA. The M6000A Mercury Analyser (Cetac Tech- before being opened. The digest was diluted with water to 5 ml prior to analysis. To estimate spike recoveries an acid nologies) CVAA instrument dedicated to the determination of mercury was used. During the analysis, mercury was reduced digestion matrix containing 20 mg l-1 of mercury was prepared by serial dilution of a 1000 mg l-1 stock solution using concen- to its zero oxidation state online by mixing with a 10% (m/v) tin chloride solution in 10% (v/v) hydrochloric acid.A small trated nitric acid. This spiked acid was then used to digest the CRM as described above and finally diluted to 5 ml with water gas–liquid separator, integral to the M6000A, was used to release a stream of mercury vapour into an argon carrier gas to produce a final spiked concentration of 2 mg l-1. In addition to the sample preparation above, a further six which passed through a magnesium perchlorate drying tube into the sample cell of the double beam AA spectrometer.larger volume digests of CRM were produced by taking 100 mg (±0.1 mg) of sample in 2.5 ml of nitric acid and diluting to After detection at 253.7 nm, by a binocular camera which compares the sample and reference beams on separate matched 25 ml. These samples were used for long-term CVAA-ICP-MS instrumental stability studies. In both cases, half the samples silicon detectors, the mercury vapour was either trapped onto solid potassium permanganate (normal operation), or coupled were digested using spiked acid with the remainder being prepared with pure or unspiked acid.directly to the ICP-MS. CVAA-ICP-MS. The interfacing of the M6000A to the Results and discussion ICP-MS was achieved by using the shortest possible length of 7 mm id Tygon tubing (approximately 80 cm), fitted with an Direct analysis using the M6000A Mercury Analyser inner sleeve of Teflon tubing, to connect the waste line of the Replicate samples of reference hair were digested in two M6000A to the back of the ICP-MS torch at the ball and groups, spiked (n=3) and unspiked (n=4).The mean blank socket joint via a Y-piece in the nebuliser flow line. When values were found to contain 0.315 (±0.065) mg l-1, with a coupled, the nebuliser flow was increased to 1.0 l min-1 and corresponding spiked value of 2.3 mg l-1. Based on a spiked the reflected power increased to about 6 W to compensate for value of 2 mg l-1 and after background subtraction, the reagent the dry sample stream.The computer software was used to spike was found to contain 1.98 mg l-1 representing a 99% select analysis points for background measurements, before recovery with an RSD of 4.3%. The mean for the unspiked and after the elemental peak and at the elemental peak CRM hair digest was found to be 1.62 mg l-1, with the spiked maximum. This also enabled the rise time to be calculated and digest giving a value of 3.67 mg g-1, representing a recovery hence the analysis time scheduling of the ICP-MS analysis.of 103% with a corresponding RSD of 3.5%. These initial experiments identified that the method Sample preparation produced acceptable recoveries for reference material and reagent spiked samples indicating a negligible loss of mercury Samples of certified human hair reference material together with a number of real hair samples from mothers who under- during the digestion process.The three acid digests of CRM shown in Table 1 gave a went normal births were prepared as follows. Approximately 20 mg of sample were accurately weighed (±1 mg) and placed mean 400±30 ng g-1 Hg in the solid compared with a certified value of 360±50 ng g-1. The mean value obtained for the in a 7 ml microvessel with 0.5 ml of concentrated nitric acid. After an initial standing period of approximately 16 h (over- CRM was therefore found to be 11% above the certified mean; however the variance obtained falls within the uncertainty night), during which time a loose cover was placed over the vessel to enable release of digest gases whilst preventing expected for the sample.In general it is recommended that a sample mass of 500 mg should be used with the CRM material contamination, the top was sealed and it was placed in a 100 ml ACV. Two microvessels were placed inside each ACV, to ensure homogeneity; however for the proposed application a typical baby hair sample size is 20 mg, which is 25 times which contained 10 ml of water.A spacer at the bottom of the ACV was used to avoid the microvessels being immersed lower than the recommended value. Accordingly the methodology used was considered to be fit for the purpose given the in the water. The outer vessel was sealed and connected to the pressure control line of the microwave oven. The water in the imposed sample size.A number of real samples analysed showed that the typical range of mercury in the hair of outer vessel allowed the microvessel to reach a high pressure without venting, by raising the gas pressure outside the micro- pregnant Hull women is between 200 and 600 ng g-1 (n=6) which compares well with other studies.5,6 vessel. Up to 6 ACVs were added to the carousel for each digestion, but only one could be pressure controlled. The outer Table 1 shows a comparison between CVAA and ICP-MS by conventional nebulisation of CRMsamples (n=3) and real hair pressure control vessel was maintained at 100 psi (1 psi= 6894.76 Pa) for up to 30 min.After cooling to room tempera- (n=6). Whilst the CRM values are acceptably close to the certified value by CVAA, the ICP-MS results were found to be ture, the inner vessels were removed and further cooled in ice Table 1 Comparison of the results for mercury determination in CRM and real hair samples by direct CVAA and ICP-MS analysis; CRM certified value 360±50 ng g-1 CVAA Conventional nebuliser ICP-MS Sample Hg in solid/ng g-1 Recovery (%) Hg in solid/ng g-1 Recovery (%) ICP-MS as a % of CVAA CRM 1 406 113 189 53 47 CRM 2 424 118 286 80 66 CRM 3 370 103 224 62 60 Mean CRM 400 (n=3) 111 233 (n=3) 65 58 Hair 1 326 148 45 Hair 2 418 259 62 Hair 3 585 344 58 Hair 4 467 290 62 Hair 5 190 131 69 Hair 6 586 289 49 Mean hair 429 244 57 128 J.Anal. At. Spectrom., 1999, 14, 127–129Table 2 Coupled CVAA-ICP-MS results for real hair samples, with control systems, introduced sample timing errors.Small errors mercury expressed for the solid sample and corresponding isotope in the timing at the beginning of a run become a major ratios calculated from raw intensities problem after about ten samples. The software control of the M6000A would ideally be via the PQVision program for the CVAA/ ICP-MS/ DiVerence best coordination. Sample ng g—1 ng g-1 (%) 200Hg5202Hg As indicated previously, one of the main reasons for 1 559 452 19.1 0.7822 developing the instrumental configuration described was to 2 362 299 17.4 0.7836 facilitate the quantitative determination of mercury and stable 3 419 360 14.1 0.7816 isotopic ratio data, which could be of value in proposed 4 94 68 27.8 0.7874 clinical studies where enriched isotopically labelled material 5 380 324 14.7 0.7827 may be used in dental work.Unfortunately, owing to the 6 169 168 0.01 0.7783 prohibitively high cost of enriched isotopic standards this 7 291 264 0.09 0.7855 aspect of the work could not be completed in this study.The Theoretical 200Hg5202Hg 0.7736 results in Table 2 do however confirm that the method used is suitable for the determination of the natural isotopic ratio 58% of their expected values. Similarly, for the real hair samples, present in the samples and oVers the potential for subsequent the ICP-MS results were 57% of the CVAA values. These results further investigations based on isotopically enriched mercury.confirm the significant losses of mercury that occur when using direct ICP-MS as reported in previous studies.2,7 Conclusions The digestion methodology described was found to be suitable Analysis by coupled CVAA-ICP-MS for small sample masses with the CVAA method giving Table 2 shows the concentration data from the coupled acceptable concentration information for mercury. Coupling CVAA-ICP-MS technique on a series of seven real hair of the CVAA to ICP-MS allowed for isotopic information to samples.Whilst the ICP-MS results show the same trends as be obtained, but a problem of long-term accumulation of the CVAA results, they appear on average to give values 13% mercury in the ICP-MS was observed owing to the sample lower than the CVAA results. This apparent diVerence in the stream being dry. Whilst stable isotopically enriched mercury mercury values is probably due to adsorption onto the inner is commercially available, it remains for the purpose of this surface of the Tygon tube, which was used in this present present study prohibitively expensive. However, the methodstudy, as no Viton tubing was readily available.This indicates ology developed indicated suitability for future tracer studies that the close coupling, with removal of the spray chamber, with labelled dental amalgam. In addition future work could has reduced the loss of mercury observed for the direct involve the interfacing of the sample introduction at a T-piece ICP-MS analysis, which gave recoveries on the order of 57%.behind the torch, which would allow the dry mercury stream A digest blank analysed between samples showed that the to be introduced into the plasma, together with an acid ICP-MS signal dropped from 1.5 ppb to a blank level of digested sample introduced via the conventional nebuliser around 0.14 ppb, reinforcing the fact that the technique has system.In this way other elements in a sample could be little carryover in the ICP-MS during a short experimental determined via the nebuliser whilst mercury is determined run. Three samples, digested with spiked acid, showed a mean using the coupled CVAA system with ICP-MS data being recovery of 108% for CVAA and 87% for ICP-MS. This obtained for isotopic ratios. With such a configuration the contrasts with conventional ICP-MS methods which use a wet nebulised stream could be made up of a gold or hydrogen nebulised sample stream and where mercury adhesion to the bromide matrix and an internal standard to sweep the mercury glass components must be minimised by addition of heavy from the internal parts of the ICP-MS, which would reduce atoms to the sample such as gold or HBr,3 which could be the long-term drift problem observed. Thus the automated sources of contamination.determination of mercury by CVAA-ICP-MS could be Results however from a long-term analysis of CRM digests, achieved simultaneously with full trace metal analysis by using the coupled technique, for which the mean of three pneumatic aspiration.determinations (i.e., 1.5 min per analysis ×3) was observed over a 4.5 h period, indicated the presence of a memory eVect, Acknowledgements apparent after 90 min, associated with the ICP-MS system. This The authors acknowledge Simon Lofthouse for practical eVect produced a gradual drift upwards of the mercury signal, assistance and discussions.The real hair samples were provided which could not be compensated for by the addition of a by Dr. S. Lindow of Hull Maternity Hospital, Kingston Upon chemical modifier, owing to the dry running of the ICP-MS. Hull, UK. We would like to thank Mr Craig Parker and Cetac Whilst the presence of a memory eVect would seriously impair Technologies (UK) Ltd. for the loan of the M6000A the determination of mercury by ICP-MS directly, the coupled instrument. approach adopted gave reliable quantitative mercury data with stable isotopic ratio information.References The coupling of the two instruments was relatively simple and solutions containing mercury introduced by the ASX 500 1 W. L. Clevenger, B. W. Smith and J. D. Winefordner, Crit. Rev. Anal. Chem., 1997, 27, 1. autosampler of the M6000A could be monitored by both 2 I. B.-A. Razagui and S. J. Haswell, J. Anal. Toxicol., 1997, 21, 149. instruments in real time. The main problem encountered was 3 J. Yoshinaga and M. Morita, J. Anal. At. Spectrom., 1997, 12, 417. with the ICP-MS PQVision software which required the instru- 4 H. T. Delves and M. J. Campbell, J. Anal. At. Spectrom., 1988, ment’s Gilson 222 autosampler to be running during the 3, 343. analysis procedure, otherwise the program halted and 5 C. G. Bruhn, A. A. Rodrý�guez, C. Barrios, V. H. Jaramillo, prompted for the next sample. The CVAA ASX 500 auto- J. Becerra, U. Gonza�lez, N. T. Gras, O. Reyes and S. Salud, J. Anal. At. Spectrom., 1994, 9, 535. sampler introduced the sample to the M6000A, which then 6 A. Oskarsson, B. J. Lagerkvist, B. Ohlin and K. Lundberg, J. Total passed into the PQ2, but the PQVision software was unaware Environ., 1994, 151, 29. of the unconventional route of sample introduction, and 7 M. J. Powell, E. S. K. Quan and D.W. Boomer, Anal. Chem., 1992, assumed that the sample was being introduced by its own 64, 2253. Gilson 222 autosampler. This problem, which arises from the lack of coordination with the two independent computer Paper 8/05105D J. Anal. At. Spectrom., 1999, 14, 127–129
ISSN:0267-9477
DOI:10.1039/a805105d
出版商:RSC
年代:1999
数据来源: RSC
|
6. |
Evaluation of atomic fluorescence and atomic absorption spectrometric techniques for the determination of arsenic in wine and beer by direct hydride generation sample introduction |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 131-135
Marta Segura,
Preview
|
|
摘要:
Evaluation of atomic fluorescence and atomic absorption spectrometric techniques for the determination of arsenic in wine and beer by direct hydride generation sample introduction† Marta Segura, Yolanda Madrid and Carmen Ca�mara* Departamento de Quý�mica Analý�tica, Facultad de Quý�micas, UCM, 28040 Madrid, Spain Received 10th July 1998, Accepted 21st September 1998 A simple and rapid flow injection-arsenic hydride generation-atomic fluorescence spectrometric (FI-HG-AFS) method has been optimized for the direct determination of arsenic in untreated samples of beer and wine.Several aspects were evaluated: concentration of HCl and NaBH4, argon and hydrogen flow-rates and eVect of oxidation state on the analytical signal. The optimum conditions selected were 6 M HCl and 0.5% (m/v) NaBH4 and 6 M HCl and 1% (m/v) NaBH4 to determine total arsenic contents of diVerent brands of beer and wine, respectively. The accuracy of the method was evaluated by comparing the results obtained with those provided for three mineralization procedures: HNO3–H2O2, H2SO4–HNO3–HClO4 and Mg(NO3)2–MgO mixtures.The results obtained showed good agreement between the conventional digestion procedures and the direct determination of arsenic in the untreated samples. The method enables the determination of arsenic to be made in untreated samples with detection limits of 0.3 and 0.5 mg L-1 for beer and wine, respectively. nitrate and nickel sulfate, have been used to circumvent analyte Introduction volatilization and matrix interferences.However, sample Arsenic is an element that has received considerable attention decomposition is required in most cases to improve sensitivity because of its high toxicity and the broad range of organs and and to achieve quantitative recovery.7 systems that are aVected in man and animals, through many Atomic fluorescence spectrometry (AFS) is a highly suitable routes.1 Arsenic is present in wines and other beverages because technique for trace analysis because of its high sensitivity, its of its wide distribution in the environment.1 Today arsenic is wide linear range of 4–6 orders of magnitude and the relatively used principally in agriculture and related areas: herbicides, simple and inexpensive equipment required.However, matrix fungicides, wood preservatives, insecticides, etc.2 Many of interference eVects seriously limit its practical use. Compared these pesticides have been banned by the Environmental with AAS, the range of applications of AFS is limited, its Protection Agency (EPA) and the use of some others is most important application being in conjunction with hydride expected to remain stable or to decline in the future.2 The generation (HG) where the analyte is separated from the arsenic content of wine and beer depends on numerous factors, matrix.8,9 This paper reports for the first time the evaluation including type of soil, climatic conditions, pollution, variety of atomic fluorescence spectrometry in conjunction with of grape, use of fertilizers and pesticides, and viticultural and FI-hydride generation for As determination in wine and beer manufacturing processes.samples. Arsine was generated directly from the wine and beer Legislation concerning arsenic contents exists in some samples. The concentrations of HCl and NaBH4 were optimcountries. In the particular case of Spain, the legislation ized for each type of sample.Accuracy, precision and selectivity specifies limits of 0.1 and 1 mg L-1 (Pb and As in combi- of the proposed method were evaluated and compared with nation) for beer3 and wine,4 respectively. those obtained by conventional HG-AAS and digestion Little has been published concerning As in wine and other procedures. beverages, but uncontaminated samples probably contain only a few mg L-1 of As. Therefore a very sensitive method is Experimental required for analyses.Traditionally, the oYcial method for determining As in wines and beer is hydride generation atomic Apparatus absorption spectrometry.5,6 However, the detection limits are Atomic absorption spectrometric measurements. A Perkin- too high to enable As to be determined at very low concen- Elmer Model 2380 atomic absorption spectrometer equipped tration levels. Further, the hydride generation technique norwith an electrodeless discharge lamp operated at 8 W from an mally requires decomposition of the sample, which is time external supply was used for all determinations.A spectral consuming and could lead to high blank values or loss of band width of 0.7 nm was selected to isolate the 193.7 nm As the analyte. line. The signals were recorded on a Perkin-Elmer Model 56 Electrothermal atomic absorption spectrometry (ETAAS) recorder set to 10 mV full-scale deflection range. should be an attractive alternative to hydride generation since the sample can be injected directly into the graphite furnace.7 Atomic fluorescence spectrometric measurements. An atomic Several chemical modifiers, such as palladium nitrate, nickel fluorescence spectrometer, PSA Analytical Model 10.033 Excalibur, was used for As determinations.The excitation source used in the instrument was an As boosted discharge †Presented at the Ninth Biennial National Atomic Spectroscopy Symposium (BNASS), Bath, UK, July 8–10, 1998. hollow cathode lamp (BD HCL, Superlamp, Photron) oper- J.Anal. At. Spectrom., 1999, 14, 131–135 131Table 1 Optimum experimental conditions ated at primary and boost currents of 27.5 and 35.0 mA, respectively. A hydrogen diVusion flame was used as the atom Parameter Value cell and an interference filter served to achieve wavelength isolation. The fluorescence signal was viewed at 90° and was NaBH4 concentration 0.5 and 1% (m/v) for beer and wine, imaged by lens onto a solar blind photomultiplier operating respectively HCl concentration 6 M at 500 V.The signal was recorded on a BBC Goerz Metrawatt Argon flow-rate 250 mL min-1 SE 120 recorder. A hygroscopic membrane dryer tube (Perma Hydrogen flow-rate 100 mL min-1 Pure, Products) was used to remove moisture from the Reagents flow-rate 2 mL min-1 gas–liquid separator. Sample injection volume 75 mL The continuous-flow and FI manifolds used to generate the arsine have been described previously. They are based on the use of a four-channel peristaltic pump (Gibson HP4), a mixing and a reaction coil (Teflon tubing, 0.5 mm id) and a U-tube 1 h, 250 °C for 2 h and 540 °C for 14 h.The resulting white gas–liquid separator. Arsine was carried to the detector by ash was left to cool at room temperature, dissolved with argon. In the FI system, the sample was injected into the concentrated HCl and then diluted to the mark with Milli-Q continuous acid flow through a six-way valve (Omnifit). water.The blank was prepared by carrying out the decomposition procedure with 10 mL of ashing aid solution. Reagents All reagents used were of analytical-reagent grade or higher Sample analysis purity and de-ionized water provided by a Milli-Q system was Arsenic was determined directly by FI-HG-AFS in wine and used. A 1000 mg L-1 AsIII standard stock solution was prebeer and in their respective digested solutions by using the pared by dissolving 0.1734 g of NaAsO2 in de-ionized water experimental conditions compiled in Table 1.Beer samples and diluting to 100 mL. The AsV stock standard solution were degassed in an ultrasonic bath prior to analysis in order (1000 mg L-1) was prepared by dissolving 0.4436 g of to improve the analytical performance of the determination. As2O5·2H2O in de-ionized water and diluting to 250 mL. The In each case, 75 mL of wine or beer or the respective digested working solutions were freshly prepared every day by diluting solution were injected into a continuous flow of 6 M HCl.The appropriate aliquots of the stock solutions. Borohydride soluresulting solution was mixed with 0.5 and 1% (m/v) NaBH4 tions of 0.5, 1, 2, 3, 4 and 10% (m/v) were prepared by solution for beer and wine, respectively. Arsine was carried by dissolving NaBH4 powder (Aldrich) in de-ionized Milli-Q argon at a flow-rate of 250 mL min-1 to the a, which water and stabilizing with 0.1% m/v NaOH solution. Solutions utilizes an argon-diluted hydrogen diVusion flame.To improve were filtered before use to eliminate turbidity. the maintenance of the flame, an auxiliary stream of H2 at The ashing aid, Mg(NO3)2–MgO, used to perform dry 100 mL min-1 was used. Analytical peaks were recorded as ashing sample treatment, was prepared by dissolving 80.0 g of peak height. Mg(NO3)2 in 200 mL of water and then adding 8.0 g of MgO. The mixture was shaken well before use. Results and discussion Sample preparation Determination of As in wine and beer by HG-AAS: the need for To validate the results obtained by direct analysis of wine and D2 background correction beer samples, several sample treatments were used to carry DiVerent preliminary studies were performed to evaluate the out sample decomposition.capabilities of the HG-AAS technique to determine As in wine and beer samples without prior mineralization. For this pur- Wet mineralization with HNO3–H2O2 mixture. A 10mL pose, the optimization of arsine generation in several ethanolic portion of sample (wine or beer) was treated with 1 mL of media (5 and 10%) and in wine and beer samples was 65% (v/v) HNO3 and 2 mL of H2O2 in a Pyrex tube placed in performed.an aluminium block and heated at 100 °C overnight. Aliquots It was observed that the amount of NaBH4 required to of H2O2 were added until the solution remained transparent generate arsine eYciently in these media was considerably and clear. The solutions were left to cool to room temperature higher than that required to generate arsine in aqueous media.and diluted with de-ionized Milli-Q water. The blank was The absorbance signal increased with NaBH4 concentration prepared by carrying out the decomposition procedure with up to 10%, where it levelled oV. Therefore, a concentration of de-ionized water instead of sample. The resulting solutions 10% m/v NaBH4 was chosen for further experiments, whereas were analysed by HG-AFS.in aqueous solution there was a much lower optimum of about 3% (m/v). The eVect of HCl in ethanolic media was similar to Wet mineralization with H2SO4–HNO3–HClO4 mixture. that in aqueous media, the optimum concentration being Sample solutions were prepared by adding 10 mL of sample 4 M HCl. (wine or beer) and 3 mL of HNO3–H2SO4–HClO4 (3+1+1) Under these conditions, significant nonselective background mixture to a Pyrex tube placed in an aluminium block and absorption was observed when arsine was generated directly heated at 100 °C.Additions of mixture were repeated until a from ethanolic media and beer and wine samples. This made completely clear and colourless solution was obtained. The it necessary to perform the analysis using simultaneous D2 solutions were left to cool to room temperature, transferred background correction to avoid errors in the determination. into a calibrated flask and diluted to the mark with de-ionized This background absorption could be attributed to the etha- Milli-Q water.The resulting solutions were analysed by nolic vapours transported to the atomizer from the hydride HG-AFS. A reagent blank was prepared in parallel. generation system. However, even with background correction, the sensitivity of the technique was very limited and no As Dry ashing with Mg(NO3)2–MgO mixture. About 10 mL of ashing aid [Mg(NO3)2–MgO] were added to 10 mL of sample signal was detected for the samples. Although this technique can be used to monitor the arsine in a 150 mL porcelain crucible.The crucible, covered with a lid, was transferred into a muZe furnace where samples were concentration in contaminated samples (with arsenic levels above the legal requirements), it is not, however, sensitive subjected to the following temperature programme: 100 °C for 132 J. Anal. At. Spectrom., 1999, 14, 131–135enough to determine this analyte directly in uncontaminated samples where the concentration level is only a few mg L-1.Determination of As in wine and beer by HG-AFS In order to achieve the required low limits of detection, the use of atomic fluorescence spectrometry in conjunction with FI-hydride generation for arsenic determination in wine and beer samples was evaluated for the first time. For this purpose, arsine was directly generated from the samples of interest. In a first approach, both samples, beer and wine, were continuously pumped into the hydride generation manifold and the arsine formed was detected by atomic fluorescence spectrometry.Under these conditions, the analytical signal did not Fig. 2 EVect of HCl concentration on fluorescence signal: (a) without reach a steady maximum value and kept increasing indefinitely an additional H2 flow and (b) with an additional H2 flow. AsIII even when experimental parameters such as HCl, NaBH4 10 mg L-1. concentration, flow-rates and pump speed were varied within a wide range of experimental values.This problem, arising raise the signal-to-noise ratio. Thus a concentration of 4 M because of the sample complexity, was overcome by introduc- HCl acid was chosen as the optimum value. ing the sample in flow injection mode by means of a six-way Initially, pure argon was used as the carrier gas and its injection valve. influence on the arsenic response is illustrated in Fig. 3. An argon flow-rate of 300–400 mL min-1 was found to be Optimization of experimental parameters. To achieve the optimum.best possible analytical performance, several experimental Although under the optimized conditions a suitable flame parameters were optimized: volume of sample injected, tetra- was generated using the hydrogen liberated from the reagents hydroborate and acid concentration, and argon and hydrogen (sodium tetrahydroborate and hydrochloric acid), the con- flow-rates. These parameters were tested for each type of ditions required to generate arsine from beer and wine sample.Fig. 1(a) shows the influence of NaBH4 concentration samples were restrictive and had to be carefully controlled to on the analytical signal for the direct analysis of wine. In the ensure the good performance of the whole process. present method, NaBH4 is used not only as a reductant but Furthermore, after five successive sample injections, the flame also as the hydrogen supply, which is necessary to sustain the was extinguished and it was diYcult to light it.argon–hydrogen flame. Consequently, in this technique, the To overcome these limitations the use of an additional concentration of NaBH4 has a large impact on the signal external flow of hydrogen was explored to broaden the range response of the hydride forming elements. For the two matrices of working conditions and improve the robustness of the tested (wine and beer) and without hydrogen as an auxiliary flame. For this purpose, an additional H2 flow was linked into gas, a minimum concentration of 1.5% NaBH4 was necessary the system and was varied within the range 100–140 mL min-1. for both in order to generate arsine and sustain the flame.As For this study, the concentrations of NaBH4 and HCl were can be seen in Fig. 1(a), the signal intensity for As initially decreased with respect to the levels optimized previously, being increases with NaBH4 concentration, but levels oV past a adjusted to 0.5% (m/v) and 0.5 M, respectively.The best result concentration of 2% NaBH4. However, at the same time the in terms of sensitivity and signal-to-noise ratio was obtained flame noise also increases, producing a deterioration in the with 100 mL min-1 H2. At the H2 flow-rate selected, the signal-to-noise ratio. A concentration of 2% NaBH4 was chosen NaBH4 and HCl concentrations were again optimized. As can for further experiments. The same behaviour was observed for be seen, the use of H2 extends the range of working conditions.beer samples. These concentration values were substantially Concentrations of 0.5% NaBH4 [Fig. 1(b)] and 2 M HCl reduced by generating the flame with an additional flow of [Fig. 2(b)] were selected as optimal to generate arsine directly H2, discussed below. from the samples. The influence of the HCl concentration in the carrier solution was investigated [Fig. 2(a)]. The HCl concentration was also EVect of oxidation state on the analytical signal critical, and a minimum of 4 M HCl was necessary to ignite and sustain the argon–hydrogen flame.Concentrations above Since the yield of AsH3 formation diVers from that of AsIII and AsV, a prereduction step is needed to overcome possible this value did not improve the analytical signal, but they did errors in total arsenic determination. Potassium iodide and cysteine solution with or without ascorbic acid have gained acceptance as the most suitable reagents for this purpose.10 Fig. 1 EVect of NaBH4 concentration on fluorescence signal: (a) without an additional H2 flow and (b) with an additional H2 flow. Fig. 3 EVect of Ar flow-rate on fluorescence signal. AsIII 10 mg L-1. AsIII 10 mg L-1. J. Anal. At. Spectrom., 1999, 14, 131–135 133Table 2 Calibration for AsIII and AsV determinations acid and reductant concentrations were changed from the levels previously optimized in an attempt to achieve identical Sample signals for the same concentrations of AsIII and AsV in beer (generation conditions) AsIII AsV and wine samples. The results (Table 2) show that in 6 M HCl, similar signal intensities for the two oxidation states of As in Wine Slope: 6.5400 Slope: 1.9800 (0.5% NaBH4, 2 M HCl ) Intercept: 22.00 Intercept: 23.40 beer were obtained.However, for the particular case of wine r2=0.9951 r2=0.9941 samples, it was necessary not only to increase the acid concen- Wine Slope: 7.6958 Slope: 7.6753 tration (6 M), but also the reductant concentration (1% m/v).(1% NaBH4, 6 M HCl) Intercept: 24.9 Intercept: 24.3 Consequently, for later applications 6 M HCl and 0.5% NaBH4 r2=0.9958 r2=0.9996 and 6 M HCl and 1% NaBH4 were selected as optimum Beer Slope: 4.3255 Slope: 1.9860 conditions to determine total As in beer and wine samples, (0.5% NaBH4, 2 M HCl) Intercept: 10.91 Intercept: 11.00 r2=0.9987 r2=0.9976 respectively, in order to avoid possible errors in the Beer Slope: 4.4876 Slope: 4.4773 determination.(0.5% NaBH4, 6 M HCl) Intercept: 10.91 Intercept: 11.02 r2=0.9997 r2=0.9987 Applications to real samples The proposed method was applied to the determination of Table 3 Slope ratios for each sample total arsenic in diVerent untreated brands of wine and beer using the experimental conditions previously optimized. In Slope ratio Slope ratio order to evaluate the selectivity of the method, several arsenic Medium medium5aqueous sample5ethanol (as AsIII ) calibration graphs were prepared in aqueous media, Aqueous 1.00 ethanolic media [10 and 5% (m/v)], beer and wine, and were Ethanolic (10%) 1.21 1.00 used to calculate the aqueous-to-ethanol slope ratio and the Red winea 1.36 1.12 aqueous-to-sample slope ratio for each type of sample. White winea 1.31 1.08 Concentrations of 5 and 10% (m/v) ethanol were chosen since Ethanolic (5%) 1.10 1.00 the ethanol content in beer and wine, respectively, is generally Beerb 1.15 1.04 similar, although not identical.The results presented in Table 3 Beer without alcohol 1.05 show that on changing from aqueous to ethanolic media the aReferred to ethanolic medium (10%). bReferred to ethanolic arsine generation eYciency slightly increases, as is evidenced medium (5%). by the aqueous-to-ethanol slope ratio values derived from the respective calibration graphs. The main matrix interference seems to originate from the ethanol content in samples of this However, the diVerence in eYciency is highly dependent on the generation media (pH, sodium tetrahydroborate concen- type. The ethanol-to-sample slope ratio had a value of around 1 in all cases.In addition, no matrix interference was observed tration, type of sample, etc.) and in some cases the addition of these reductants is not necessary. in the determination of As in beer samples without alcohol, where similar slopes were obtained for aqueous and beer The eVect of arsenic oxidation state on the analytical signal was evaluated by adding increasing concentrations of AsIII and calibration graphs.All these results indicated that the proposed method seems to be highly selective and that the main AsV to the samples. Arsine was later generated under the optimum conditions previously established (0.5% NaBH4, 2 M interference is caused by the amount of ethanol; however, it is recommended that the standard additions be used to HCl) and determined. DiVerent arsine generation eYciencies (resulting in diVerent slope values) were obtained for AsIII and determine As.The accuracy of the method was tested by comparing the AsV in the two types of samples, which made it mandatory to perform a prereduction step or to change the generation results obtained with those provided for three conventional mineralization procedures: wet mineralization with an conditions. The addition of potassium iodide up to a concentration of 20% (m/v) did not improve the sensitivity for AsV H2SO4–HNO3–HClO4 mixture, wet mineralization with an HNO3–H2O2 mixture and dry ashing with an in any case.The fact that this reductant was not able to transform AsV into AsIII can be explained by the low solubility Mg(NO3)2–MgO mixture. The results obtained are compiled in Table 4. Comparison by the F-test of the results provided of this compound in ethanolic medium and/or by the matrix composition, which may have modified the potential values of by the two methods (without and with mineralization) revealed no significant diVerences at the 95% confidence level.Also, the implied redox systems. The eVect of the arsenic oxidation state can be eliminated the results obtained when the arsine generation eYciency did not depend on the oxidation state were similar to those or decreased by changing the generation conditions. Thus, the Table 4 Arsenic concentration in wine and beer samples (mg L-1); n=6 determinations Without mineralization With mineralization 6 M HCl, 1% NaBH4 (wine) Sample 6 M HCl, 0.5% NaBH4 (beer) 2 M HCl, 0.5% NaBH4 HNO3–H2O2 H2SO4–HNO3–HClO4 Ash digestion Montenoble table wine 5.2±0.1 5.3±0.2 4.8±0.4 — 5.0±0.2 Cumbre de Gredos table wine 3.7±0.3 3.2±0.2 3.2±0.2 — 3.6±0.2 Don Opas table wine 3.1±0.2 2.8±0.3 3.0±0.1 — — Oporto 2.8±0.1 — — — — Vin�a Zaco (Rioja) 2.6±0.2 — — — — Rioja Berberana 2.0±0.2 — — — — Los Molinos (table wine, white) 3.8±0.1 — — — — Mahou beer 2.5±0.2 2.3±0.2 — 2.2±0.1 2.0±0.2 Cruz Campo beer 2.3±0.1 2.0±0.2 — — — Buckler beer (without alcohol ) 2.5±0.2 2.0±0.3 — 2.0±0.2 — Damn beer (without alcohol ) 5.3±0.1 4.8±0.2 — — — 134 J.Anal. At. Spectrom., 1999, 14, 131–135obtained using 2 M HCl and 0.5% (m/v) NaBH4, which Acknowledgements provided a diVerent eYciency. This fact may indicate that The authors thank the DGICYT for financial support under arsenic is mostly present in wine and beer as AsIII. contract PB 95-0366 C01–C02. The detection and quantification limits calculated according to the IUPAC guidelines were, for the direct determination of As in beer, 0.3 and 1.0 mg L-1, respectively, and in wine, 0.5 References and 1.6 mg L-1, respectively. The precision of the method for 1 Commission of Codex Alimentarius, FAO/OMS, Arsenic ten replicates on each of the diVerent samples of beer and Document., CX/FAC 97/23 Part Y. wine analysed ranged between 3 and 8%. All these results 2 S. Ishiguro, Appl. Organomet. Chem., 1992, 6, 323. make the proposed method highly suitable for determining As 3 Spanish Royal Decree 1456/81. in uncontaminated beer and wine samples. 4 Spanish Royal Decree 3024/73. 5 M. L. Cervera, A. Navarro, R. Montero and R. Catala�, J. Assoc. OV. Anal. Chem., 1989, 72, 282. 6 C. Balya-Santos and A. Gonzalez-Portal, Talanta, 1992, 39, 329. 7 B. J. Kildahl and W. Lund, Fresenius’ J. Anal. Chem., 1996, 354, Conclusions 93. 8 W. T. Corns, P. B. Stocl, L. Ebdon and S. J. Hill, J. Anal. At. The proposed method is fast, accurate and sensitive and can Spectrom., 1993, 8, 71. be used to quantify As at its naturally occurring level in wine 9 S. J. Hill, J. B. Dawson,W. J. Price, I. L. Shuttler and J. F. Tyson, and beer without need for previous treatment. This makes it J. Anal. At. Spectrom., 1996, 11, 281R. highly promising for monitoring As in these samples and 10 A. Lopez, R. Torralba, M. A. Palacios and C. Ca�mara, Talanta, opens new pathways for incorporating atomic fluorescence 1992, 39, 1343. spectrometry into the range of techniques used in routine analyses. Paper 8/05391J J. Anal. At. Spectrom., 1999, 14, 131–135 1
ISSN:0267-9477
DOI:10.1039/a805391j
出版商:RSC
年代:1999
数据来源: RSC
|
7. |
The future of atomic absorption spectrometry: a continuum source with a charge coupled array detector† |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 137-146
James M. Harnly,
Preview
|
|
摘要:
The future of atomic absorption spectrometry: a continuum source with a charge coupled array detector† Plenary Lecture James M. Harnly US Department of Agriculture, Agriculture Research Service, Beltsville Human Nutrition Research Center, Food Composition Laboratory, Building 161, BARC-East, Beltsville, MD 20705, USA Received 29th September 1998, Accepted 17th November 1998 Continuum source atomic absorption spectrometry (CS-AAS) has made impressive progress in the last 5 years thanks to the availability of high resolution e� chelle spectrometers and solid state array detectors.With these new spectrometers and detectors, the capabilities of CS-AAS exceed those of conventional, line source-AAS (LS-AAS). For CS-AAS, absorbances are more accurate (corrected for stray radiation and non-specific broadband background absorption and integrated with respect to height in the furnace), detection limits average a factor of 2 lower, calibration ranges are a factor of 1000 greater, multi-wavelength data are available for correction of spectral interferences, sensitivity is a powerful quality assurance measure since it is independent of all instrument parameters except atomization temperature and, of course, multi-element detection is possible. The future appears bright for CS-AAS.Whereas, previously, CS-AAS was striving for parity with LS-AAS, it is now reasonable to state that it is CS-AAS which is setting the standard. Substitution of a continuum source for HCLs, without Introduction changing the rest of the instrument, is not a reasonable Continuum source atomic absorption spectrometry (CS-AAS) approach.The instability of the most intense continuum has long appealed to the spectroscopic community because of sources, xenon arc lamps, gives noisy baselines and poor the potential for simultaneous multi-element AAS determi- detection limits. Medium resolution monochromators, that are nations, a shortcoming of conventional line source AAS ideal for isolating HCL emission lines, provide a spectral (LS-AAS). Unfortunately, owing to limitations of the source, bandwidth that is too large for use with a continuum source.CS-AAS has failed for many years to compete with LS-AAS The large spectral bandwidth results in poor sensitivity and with respect to detection limits. Today, with the high radiation specificity, non-linear calibration curves and greater susceptithroughput and spectral resolution of e� chelle spectrometers, bility to broadband background interferences. In addition, the the multi-wavelength detection capability, low read noise and intensity of most continuum sources decreases dramatically high quantum eYciency of charge coupled detectors (CCDs) below 280 nm.Consequently, the use of a continuum source and the high speed data acquisition capabilities of modern for AAS requires the redesign of the whole instrument. computers, CS-AAS has surpassed the analytical capabilities As shown in Table 1, a variety of instrumental designs have of LS-AAS.These advantages will be explained in detail in been explored for CS-AAS. The three main challenges were this review. Although it takes time for the implementation of to obtain sensitivities, detection limits and calibration ranges new concepts, especially when it requires a considerable capital comparable to LS-AAS. Sensitivity was initially enhanced by investment, CS-AAS appears ripe for development.It seems the use of multipass absorption cells.3,4 It soon became obvilikely that if AAS was being developed today for the first ous, however, that the best approach to recovering the lost time, it would be with a continuum source. sensitivity was the use of a high resolution spectrometer AAS, as first described by Walsh1 and Alkemade and (interferometers and e� chelles)7–14 to provide the narrow analyt- Milatz,2 derived its success from the use of hollow cathode ical bandwidth previously supplied by the HCL.Obtaining lamps (HCLs) as the radiation source. These lamps, with their detection limits equal to those for LS-AAS required minimizing stable and narrow emission lines, oVered high analyte speci- the flicker noise of the xenon arc lamp. This was accomplished ficity, excellent detection limits (for the time) and linear using wavelength modulation with phase sensitive deteccalibration curves (2.5–3.0 orders of magnitude of concen- tion.5–7,10–14 Wavelength modulation has been implemented tration).Unfortunately, these lamps are not suitable as multi- using quartz refractor plates,5,6,10,11 oscillating e� talons7 and, element sources. Multi-element HCLs are restricted to compat- ultimately, array detectors.12–14 The linearity of the calibration ible elements and are always less intense than the single- curves is determined primarily by the spectral bandwidth of element lamps. Combinations of single- and multi-element the analytical measurement.Thus, the e� chelles and interfer- HCLs suVer from loss of intensity due to the necessary beam ometers have provided the best linearity,8,10 although the splitters/combiners. Consequently, although AAS is a powerful oscillating e� talon produced an unusual, multi-humped calianalytical tool, it has always remained a single-element bration curve owing to the narrowness of the spectral bandtechnique. pass. It should be noted that all the instruments described in Table 1, with one exception, are single-element designs.The only functional, multi-element instrument was that described †Presented at the Ninth Biennial National Atomic Spectroscopy Symposium (BNASS), Bath, UK, July 8–10, 1998. by Harnly et al.11 in 1979. J. Anal. At. Spectrom., 1999, 14, 137–146 137Table 1 Research on continuum source AAS Year Researchers Instrument Design Reference 1966 Fassel, Mossotti, Grossman Long cell paths 3 and Knisely 1967 McGee and Winefordner Long cell paths 4 1968 Snelleman Wavelength modulation 5 1972 Elser and Winefordner Double modulation (wavelength 6 modulation and optical chopping) 1972 Nitis, Svoboda and Fabry–Perot interferometer with 7 Winefordner oscillating etalon 1973 Veillon and Merchant Fabry–Perot interferometer 8 1974 Keliher and Wholes E� chelle monochromator 9 1976 Zander, O’Haver and E� chelle monochromator and 10 Keliher wavelength modulation 1979 Harnly, O’Haver, Golden E� chelle polychromator, wavelength 11 and Wolf modulation and computerized data acquisition (SIMAAC) 1993 Harnly E� chelle monochromator and linear 12 photodiode array detection 1997 Harnly, Smith, Wichems, E� chelle polychromator and segmented, 13 Ivaldi, Lundberg and Radziuk linear charge coupled array detectors 1998 Harnly, Fields and Schuetz E� chelle monochromator and two- 14 dimensional, thinned, back-illuminated charge coupled array detector Wavelength modulation is the heart of CS-AAS.Intensities data. A frame consists of all the data obtained from an array following an exposure for a predetermined period of time. A on and oV the analytical line are ratioed to correct for broad band shifts in the spectral intensity, whether caused by the frame may consist of data from all the pixels or select groups of pixels (sub-arrays) covering specific wavelength regions of lamp flicker or non-specific background absorption. Wavelength modulation is successful, however, only if it is interest.The frame rate of an array detector can be viewed as equivalent to the modulation frequency of the mechanical performed at a frequency greater than that of the intensity fluctuation. With a xenon arc lamp, a modulation frequency modulation devices. Since each frame consists of intensities measured over the whole wavelength region of interest, high of 50–60 Hz is needed to minimize the source fluctuation noise.15 The development of electrothermal atomization did frame rates are not necessary for accurate correction of intensity fluctuation.High frame rates, however, are still not significantly influence the data acquisition rate for CS-AAS. The 60 Hz data rate dictated by the flicker noise of necessary to characterize accurately the rapid, transient signals which result from furnace atomization. Intensities mp was suitable for furnace atomization.16 Prior to 1993, wavelength modulation was implemented converted to absorbance and integration must take place in the absorbance domain to ensure linearity with respect to mechanically and the resultant signal was detected using a lock-in amplifier.10,11 The mass of the moving mechanical concentration.Frame rates of at least 60 Hz are required for furnace atomization and even higher rates are preferred for device (the torque motor on which the quartz refractor plate was mounted) limited the modulation frequency.Frequencies the newer furnaces with heating rates of greater than 1000 °C s-1. above 60 Hz were not possible without sacrificing the signalto- noise ratio (S/N). Detection with lock-in amplifiers pro- The current state-of-the-art for CS-AAS is embodied in the latest instrument listed in Table 1, developed by Harnly et al.14 vided signals (I0-I ) that were proportional to absorbance at low concentrations. With respect to data processing, any of This design, which will be described in detail, is a singleelement instrument. Its characteristics, detailed in Table 2, the earlier instruments in Table 1 could have been adapted to multi-element detection using parallel analog circuits.illustrate the potential capabilities of CS-AAS using the latest solid state detector technology. The low pixel read noise results Coupling of a computer to CS-AAS in 197911 permitted the use of an exotic modulation waveform and high-speed data in photon shot noise being the dominant noise source. As a consequence, the best S/N can be achieved with the highest acquisition for all elements.SIMAAC acquired 20 intensities per modulation cycle for 56 cycles per second for 16 elements, resolution. The use of a high-resolution e� chelle spectrometer restricts the spectral width of each pixel, allowing high sensi- a total data acquisition frequency of 17.9 kHz. The combination of the unusual waveform and high data rate resulted in intensity measurements at discrete wavelengths.The 20 Table 2 Spectrometer characteristics intensity measurements acquired during a single sweep across E�chellea Dectectorb the absorption profile were used to compute a family of absorbances of varying sensitivity. This approach allowed 750 mm focal length Split frame transfer array detection of 16 elements, improvements in the S/N at low 63° 26¾ blaze angle 80×80 pixels concentrations and extension of the calibration curve linearity 46×96 mm grating 18×18 mm pixels to 4–5 orders of magnitude. 79 grooves mm-1 2.5 mm channel stop In the 1990s, use of solid state array detectors permitted 25 mm entrance slitwidth Thinned 500 mm entrance slit height Back-illuminated simultaneous multi-wavelength detection and eliminated the Spectral bandwidth: UV antireflection coating need for mechanical modulation.12–14 Multi-wavelength detec- 2.1 pm at 200 nm 100 000 e- charge capacity tion provides the equivalent of infinitely fast wavelength 7.0 pm at 580 nm 70 Hz frame rate modulation (on- and oV-line intensities are measured simul- Luminosity at 200 nm: 25 e- read noise taneously) provided that all the pixels have a common exposure 0.013 mm2/nm 50% quantum eYciency at 200 nm time, i.e., the exposure start and stop times are the same.The aRef. 19. bRef. 20. array detectors compute absorbances from a single ‘frame’ of 138 J. Anal. At. Spectrom., 1999, 14, 137–146practice, the data from the camera in Table 2 had 80 columns and 80 rows of pixels.The conversion of the intensities in a single frame to absorbance is relatively simple when a computer is available to do the data processing. In Fig. 1, linear arrays of pixels running parallel to the wavelength axis will be called rows and linear arrays perpendicular to the wavelength axis will be called columns. Simplistically, the intensities at the top and bottom of each column (i.e., intensities between orders) are subtracted from the intensities in the middle of the column to correct for stray light.Intensities at the ends of each row are used to compute absorbances for each pixel in the middle of the row. Absorbances from pixels in individual columns are summed to provide height-integrated absorbances. The heightintegrated absorbances are summed to provide a wavelengthintegrated absorbance. Finally, the height and wavelength inte- Fig. 1 Simulated data showing a short wavelength section of one of grated absorbances from each frame are summed to provide the dispersed orders of the e� chelle spectrometer with a slit height of 500 mm and an absorption profile with an absorbance of 0.5 and a a time integrated absorbance.full width at half-height of 2 pixels. Pixels in black above and below Mathematically, the height, wavelength and time integrated the order are used to correct for stray light. Pixels in black on the absorbance, AInt, can be expressed as order to each side of the absorption profile are used to correct to broadband background absorption interferences. AInt=tf .frame lp . row hp . column log(I0,r/Ii,r) (1) where tf, lp and hp are normalization constants (tf is the time tivity and permitting the acquisition of detailed spectral inforbetween frame reads in seconds, lp is the spectral pixel width mation for the surrounding wavelength region. The instability in picometers and hp is the pixel height in micrometers), I0,r is of the continuum source and non-specific background absorpthe reference intensity for row r obtained from the average of tion are eliminated by simultaneous multiwavelength detection pixel intensities at the ends of the row and Ii,r is intensity of with the array detector.In addition, the high throughput of pixel i in row r. The reference pixels must be predetermined the e� chelle and the high quantum eYciency of the array by the analyst. The S/N is quadratically dependent on the detector provide increased intensity in the far UV region number of reference pixels; hence the more used the better, (below 240 nm).The vertical dimension of the array (when although there are diminishing returns as the number grows coupled with an appropriate optical arrangement) can provide large. The normalized time, wavelength and height integrated absorbance information as a function of height in the furnace absorbance will have units of mm pm s.The mass necessary to and can measure stray light intensities between orders. provide an integrated absorbance of 0.0044 pm s (integrated The instrument described in Table 2 provides analytical with respect to wavelength and time) was previously defined information that is unique to the field of atomic spectrometry. as the intrinsic mass. The mass necessary to provide an A simulation of a typical frame of data is shown in Fig. 1. absorbance of 0.0044 mm pm s (with the addition of height From these data, it is possible to compute absorbances that integration) has not yet been defined.are stray light corrected, background corrected and integrated Fig. 2 and 3 present results obtained for the atomization of with respect to wavelength, height in the furnace and time. In 250 pg of Cu (324.7 nm). The intensities shown in Fig. 2 were addition, the data are suitable for resolving overlapping specobtained by averaging the data for four frames near the time tral interferences, although this area is largely unexplored for of the peak maximum.These data were obtained with an absorption spectrometry. In theory, after normalization, the entrance slit height of 100 mm (compared with the 500 mm computed absorbances are independent of the characteristics height used in Fig. 1 for the simulated data) so the order width of the continuum source, monochromator and solid state is much smaller. Fig. 3 provides the computed absorbances detector.The analytical sensitivity will depend only on the physical characteristics of the furnace (diameter and length) and the accuracy and repeatability of the temperature program. These analytical characteristics now make L’vov’s concept of absolute analysis17,18 a realizable possibility. Computing absorbance The simulated data in Fig. 1 shows a band of elevated intensities through the frame parallel to the wavelength axis. This band is a single order of the e� chelle viewed over a short wavelength interval, typically less than 1 nm.In Fig.1, intensities above and below the band (between orders) are also seen. Stray radiation and oVsets in the electronic circuitry determine the magnitude of the intensities between orders. The height of the order above the between order intensity is determined by the source intensity and the order being viewed and fluctuates with time. The width of the band (on the order axis) will depend on the entrance slit height. The dip in the middle of the band is the analyte absorption and the depth will vary with analyte concentration.Additional dips in the band may be seen if non-analyte absorption, or spectral line Fig. 2 E� chelle order containing absorption profile for 250 pg of Cu interferences, occur. The simulated data in Fig. 1 are con- (324.7 nm) obtained with an entrance slit 25 mm wide and 100 mm high. Large vertical oVset is electronic and not stray light. structed with only 22 columns and 54 rows of pixels.In J. Anal. At. Spectrom., 1999, 14, 137–146 139model and the fact that diVerent atomization temperatures were used, the agreement between the intrinsic mass for LS-AAS and CS-AAS is very reasonable. The lack of dependence of CS-AAS on the source and detection parameters places new emphasis on the performance of the furnace. Absorbances for any CS-AAS instrument can be directly compared and absorbance will have a fixed relationship to concentration.The only variable that can cause a change in sensitivity is the furnace. The number of atoms in the light path is dependent on the temperature, the furnace length and matrix interferences. A two-step furnace24 will allow better control of two of these variables; temperature and matrix interferences. Since a two-step furnace does not require a rapid temperature ramp, accurate and reproducible temperatures should be easier to achieve. Volatilization of the sample into a constant temperature should alleviate many of the chemical interferences arising from the sample matrix.Only the furnace length would have to be standardized to obtain a Fig. 3 Absorbance map for the same data as shown in Fig. 2 (250 pg uniform response between diVerent instruments. of Cu). Absorbance was computed using the intensities from the first and last column (perpendicular to wavelength axis) as reference Noise characteristics intensities. There was no stray light correction for these data.An instrument composed of a continuum source and an array detector can be expected to have three major noise sources: for each pixel in Fig. 2. For the purpose of illustration, the fluctuation and shot noise from the xenon arc lamp and read absorbances in Fig. 3 were computed without stray light noise from the array detector. correction and using the average of the first and last pixel in Fluctuation noise is inherently eliminated using an array each row as the reference intensity.Stray light correction was detector provided that the pixels have a common exposure omitted so that the absorbances for the pixels between orders time. More specifically, this means the exposure start and stop (i.e., at the top and bottom of each column) could be computed. times for each row of pixels (parallel to the wavelength axis) With stray light correction, pixels between orders will have must be the same. If an LPDA is read sequentially, some of intensities randomly distributed around zero.As a result, the the fluctuation noise will be retained.12 A typical exposure logarithm of the ratio for many of these values will be consists of an integration period followed by a sequential read undefined and the rest will have a large variance. In routine period. Each pixel, although integrating for the same length practice, absorbances are only computed for pixels on the of time, will not be exposed to the same radiation (since the order, i.e., pixels whose intensity will be significantly greater starting and stopping times are not the same) and can be than zero after stray light correction.subject to diVerent fluctuations of the source. The longer the In theory, the height, wavelength and time integrated integration period, with respect to the read period, the more absorbance is independent of the characteristics of the source, the fluctuation noise will be reduced. Blocking exposure of the disperser and detector.Absorbances computed as described array during the read period will eliminate fluctuation noise12 above are corrected for stray light and broadband background by establishing a common starting and stopping time for each absorption and are constant for a constant number of absorbpixel for the integration interval. ing atoms in the absorption cell (furnace). The source intensity Frame transfer devices, which shift the data from the image and fluctuation, the monochromators throughput and spectral array to unexposed storage arrays prior to reading, will also bandwidth and the detectors pixel count, pixel size and quaneliminate fluctuation noise, provided the shifts are perpendicu- tum eYciency will not alter the computed absorbance.The lar to the wavelength axis.13,14 The split frame transfer camera noise component will increase as the intensity decreases, but used in this study shifts 40 rows upward and 40 rows downward the mean absorbance will not change.In addition, wavelength from the image arrays to the storage arrays. Data from the and time integration also mean that the integrated absorbance storage arrays are then read systematically during the next is not dependent on the spectral distribution of the absorption integration period. Since the shifts are perpendicular to the coeYcient (i.e., changes in width due to collisional broadening wavelength axis, the integration period for each row is diVerent at higher pressures or changes in distribution due to diVerences but the exposure period for each pixel in a row is common.in the isotopic composition) and is not dependent on the Thus, absorbances computed for each row or absorbances appearance function of the analyte (provided that the loss computed from intensities summed vertically (binned) will function is constant). eliminate fluctuation noise. In general, shifts perpendicular to Table 3 lists the intrinsic masses determined for a variety of the wavelength axis allow the elimination of fluctuation noise CS-AAS instruments over a 10 year interval.The spectral while shifts parallel to the wavelength axis will retain some of width of a pixel (at 196 nm) varies from 1.9 to 100 pm. For the fluctuation noise. some elements, the data in this table should show better Minimizing the read noise of a pixel is important for agreement. This table, however, represents data put together obtaining the best possible detection limits with the highest in retrospect and was not part of a systematic study.The resolution. It was shown, using propagation of errors,12 that biggest source of variation in intrinsic mass is the furnace the absorbance noise, sA, for the read noise limited case was atomization. In each case, an HGA 500 graphite furnace atomizer was used but the atomization temperatures between sA= 0.43sreadÓn+1 I (2) experiments varied by as much as 500 °C.The estimated intrinsic mass for LS-AAS is also listed. The estimated intrinsic mass was obtained by correcting the characteristic mass for where sread is the read noise for a pixel, n is the number of pixels needed to cover the absorption profile (and also the the linewidth ratio of the HCL and the absorption profile using a simple triangle model and assuming absorption over number of reference pixels) and I is the intensity read from each pixel. Eqn. (2) is very useful because it is directly pro- the entire absorption profile.Considering the simplicity of the 140 J. Anal. At. Spectrom., 1999, 14, 137–146Table 3 Intrinsic massa LS-AAS mi for CS-AAS E� chelled LPDA E� chellee LPDA H-20f E� chelleg SCD Wavelength/ (25 mm)h (50 mm)h LPDA (25 mm)h Element nm m0 b mi c (50 mm)h As 193.7 17 9.2 9.7 9.9 – 10.5 Se 1986.0 30 10 11 10 – 12 Zn 213.9 0.40 0.50 1.4 0.17 Pb 217.0 5.0 1.9 1.9 1.8 2.1 3.0 Sb 217.6 38 15 – – – 10 Sn 224.6 20 10 9.2 9.4 – – Cd 228.8 0.35 0.14 0.27 0.19 0.23 0.29 Ni 232.0 13 5.5 8.0 8.4 6.4 – Co 240.7 6.0 2.4 2.3 2.3 – – Fe 248.3 5.0 1.9 1.6 1.8 – – Tl 276.8 10 4.0 – – – 11 Mn 279.5 2.0 0.61 0.83 0.88 – 1.4 Pb 283.3 11 4.2 3.9 3.8 – 6.7 Cu 324.7 4.0 1.0 1.00 1.1 – – Ag 328.1 – – 0.4 – – – Spectral width (pm) 1.9 3.8 100 6.1 of Pixel at 196 nm aAll values determined with HGA-500 graphite furnace atomizer (Perkin-Elmer, Norwalk, CT, USA).bCharacteristic mass.21 cComputed from m0 as described in ref. 22. dSpectraspan III e� chelle (Spectrametrics) with 256 pixel LPDA and 25 mm pixel width.22 eSpectraspan III e� chelle (Spectrametrics) with 128 pixel LPDA and 50 mm pixel width.22 fH-20 monochromator with 128 pixel LPDA and 50 mm pixel width.23 gOptima e� chelle (Perkin-Elmer) with 256 pixel LPDA and 25 mm pixel width.13 hPhysical width of pixel. Table 4 CS-AAS detection limits portional to the detection limit. The previous section demonstrated that the normalized, integrated absorbance was Element Wave- HGA-500 furnace THGA furnace independent of the source and detector characteristics.length/ Consequently, the detection limit is directly proportional to nm LSa CS-LPDAb CS-SCDc LSd CS-DEMONe the absorbance noise. As 193.7 20 28 12 6 4 Eqn. (2) shows that the absorbance noise will grow smaller Se 196.0 30 50 16 9 13 as the intensity increases or the read noise decreases. This Zn 213.9 1 2 0.1 0.4 0.2 equation also indicates that the absorbance noise will decrease Pb 217.0 10 6 4 4 – with increasing spectral bandwidth.For example, if the Sb 217.6 15 – 8 4 2.5 entrance slit width is doubled, the intensity will double, the Bi 223.1 6 – 5 – – number of pixels necessary to cover the profile will double Sn 224.6 20 26 – 10 – Cd 228.8 0.4 0.4 0.07 0.1 0.1 and sA will decrease by Ó2. The LPDA used previously with Ni 232.0 10 11 – 8 – CS-AAS12 had a read noise of about 3000 e-. As a result, Be 234.9 1 – – 0.1 0.08 read noise was dominant at all intensity levels.The best Co 240.6 2 4 – 4 – detection limits (Table 4) were obtained with the largest Fe 248.3 2 2 – 0.8 0.6 entrance slit width of the e� chelle, 500 mm. At the time, these Si 251.6 40 – – 15 6 detection limits were the best ever achieved for CS-AAS, but Tl 276.8 10 – 1 9 3 Mn 279.5 21 0.5 0.2 0.6 0.3 they precluded operation in the high-resolution mode. Pb 283.3 5 0.9 0.4 4 1 The photon shot noise limited case, the ideal case, is achieved Al 309.3 4 – – 3 0.8 if the fluctuation noise is eliminated and the read noise is low.Mo 313.3 4 – – 1 2 A high quality CCD will typically have a read noise of less Cu 324.7 1.0 0.6 – 4 1 than 25 e-. Consequently, all but the lowest intensities Ag 328.1 0.5 – – 0.4 0.2 (<625 e-) will be shot noise limited. The absorbance noise Cr 357.9 1 – – 0.4 0.8 for the shot noise limited case is aModel 5000 (Perkin-Elmer).21 bCS with linear photodiode array (LPDA) detector.22 cCS with segmented charge coupled array detector (SCD) of Optima (Perkin-Elmer).13 dSIMAA 6000 (Perkin-Elmer).25 sA= 0.43ÓIÓn+1 I = 0.43Ón+1 ÓI (3) eCS with double e� chelle monochromator (DEMON).26 since sI=ÓI.This equation shows that, for the shot noise limited case, the absorbance noise is independent of the spectral bandwidth. Consider again the case where the entrance slit more intense sources, with little success. More signifcant improvements have been made in increasing the luminosity of width is doubled.As stated previously, both n and I will also double. With eqn. (2), sA is reduced by Ó2. With eqn. (3), the spectrometer and quantum eYciency of the detectors. Table 2 shows that the two-dimensional array is thinned and however, sA remains the same. Consequently, the use of a narrow slit width to maximize spectral resolution does not back illuminated. The result is a quantum eYciency of 50% at 200 nm. degrade the detection limit.Eqn. (2) shows that if I, the level of radiation striking each Table 4 presents detection limits for some of the most recent CS-AAS instruments. A comparison of eqns. (2) and (3) pixel, increases without opening the entrance slit width, the absorbance noise will decrease. Three means of increasing I shows why better detection limits are achieved in the far UV with a CCD array. For the LPDA described above (read noise are to increase the source output, increase the transmission eYciency of the spectrometer and increase the detection about 3000 e-), the detected intensity must be 9×106 e- in order for the shot noise to equal the read noise.For the same eYciency. Numerous attempts have been made to develop J. Anal. At. Spectrom., 1999, 14, 137–146 141detector, elements lying between 190 and 230 nm had intensit- matrix, it seems likely that they arose from molecular bands of NO. The non-analyte absorption around the As and Se ies ranging from 3.8×105 to 3.2×106 e- for a 20 ms integration interval.Thus, the predicted absorbance noise for the lines could be both spectrally and temporally resolved. The non-analyte absorption disappeared when the elements were shot noise limited case ranged from 4.5 (at 193.7 nm) to 1.6 (at 232.0 nm) times lower. The results in Table 4 show that determined in a Pd(NO3)2 chemical modifier. Selenium (196.0 nm) was the only element of those examined for which the measured detection limits for the segmented CCD detector (read noise 15 e-) ranged from equivalent to a factor of 3.1 a Pd line was observed.The only consequence of the nonanalyte absorption in the nitric acid matrix is to emphasize times lower. The improved detection limits of the CS-double e� chelle the necessity for choosing the oV-line reference pixels with care. Obviously, this same problem is of much greater concern monochromator (CS-DEMON) instrument22 can be attributed to better imaging of the lamp on the entrance slit of the for oV-line background correction for emission spectroscopy because the operating temperatures are much higher and the spectrometer.CS-DEMON used a small, xenon short-arc lamp with a well defined bright spot at the tip of the cathode. This spectra are much more complex. Schuetz26 has reported much more complex structured bright spot was easier to image on the entrance slit than that of lamps employing the reflector as an integral part of the background spectra around Cd (228.8 nm), Pb (217.0 nm) and Se (196.0 nm) when using ammonium dihydrogenphosphate lamp housing.The bright spot provided higher UV intensities and slightly better detection limits. and magnesium nitrate as a chemical modifier. An obvious solution is to not use this particular modifier, but the more general question is still of interest: can interference from Spectral resolution complex line structures be corrected? In his thesis, Schuetz used spectral subtraction with some success.This approach, The combination of high S/N with high resolution makes CS-AAS an ideal instrument for examining spectral line inter- however, requires matching the sample matrix qualitatively and quantitatively. Initial results suggest that AA spectra are ferences. LS-AAS only has emission intensity over the narrow spectral width of the HCL emission line. It is impossible to be suYciently uncongested to allow the more empirical approach of reconstructing the spectra with modeled peaks.28 At this sure what absorption lines lie just outside the emission line width.Absorption interferences that fall within the mono- time, more study is necessary. The major point, however, is that there is suYcient information available to allow correction chromator’s spectral bandwidth (for background correction with a secondary, continuum source) or are shifted within the for spectral line interferences. The only question is how much eVort is justified in making the correction an of emission line bandwidth by the magnetic field of Zeeman AAS, can only be inferred.With CS-AAS, however, it is now automation that can be achieved. possible to inspect the spectral regions around the absorption profile in a manner analogous to that used for emission Spatial resolution spectroscopy. Fig. 4 shows the spectra in the region of the Se (196.0 nm) The columns of pixels perpendicular to the wavelength axis record intensities as a function of the order height.The order absorption line as a function of time.13 In this case, Pd(NO3)2 was used as a chemical modifier so the mass of Pd (5 g) was width, determined by the height of the entrance slit, is 500 mm. With an appropriate optical arrangement,17 the order width 2500 times that of Se (2 ng). It can be seen that after 4 s of atomization the Pd (peak B) was still present, whereas the Se reflects intensities as a function of height in the furnace.In eVect, the array detector (with pixels 20 mm high) divides the (peak A) was relatively short lived. Vacuum wavelength tables from the National Institute of Standards and Technology27 radiation transmitted through the 6 mm high furnace into 25 individual sections. Separate absorbances can be computed show the presence of a weak Pd line (196.011 nm) just 15 pm from the Se line (196.026 nm). The two peaks are fully resolved for each section and summed to provide a furnace-height integrated absorbance that is linear with respect to with a spectral bandwidth of 3 pm per pixel.Similar data for As, Bi, Cd, Mn, Pb, Sb, Se, Tl and Zn in concentration. Gilmutdinov and co-workers29 have shown that there is a a 5% v/v nitric acid matrix showed some non-analyte absorption around the As and Se lines but none around the wave- non-uniform distribution of source intensity and analyte atoms in the furnace. The lateral distribution of atoms tends to be lengths of the other elements.13 The source of the non-analyte absorption was not identified although, with a nitric acid symmetrical and relatively homogeneous at any height compared with the vertical distribution which is asymmetric and non-homogeneous owing to the dosing hole at the top of the furnace. The non-homogeneous distribution of atoms produces non-uniform transmitted intensities that, with integration in the intensity domain, can result in a non-linear response with respect to the analyte concentration.Gilmutdinov and co-workers speculated that, with complex sample matrices, the absorbances computed for LS-AAS are significantly diVerent for standards and samples owing to the non-homogeneity of the spatial distribution of the analyte atoms. McNally and Holcombe30 demonstrated a non-uniform height distribution for Cu atoms in a pure standard. With the furnace height divided into nine subsections, they measured absorbances at the bottom and the top of 0.39 and 0.30, respectively, at the time of the peak maximum.This diVerence of 0.09 is 23% of the signal at the bottom of the furnace at the time of the peak maximum. In Fig. 5, the diVerence in working range arise from the finite width of the HCL emission line (the HCL line is not monochromatic, just 3–5 times narrower than the absorption profile at atmospheric pressure) and stray light. With CS-AAS, there is no theoretical limit to the calibration range, only the practical limits imposed by the size of the array, the increasing possibility of spectral interferences and the ability to clean the furnace between atomizations.The shapes of the calibration curves were predicted theoretically in the original work by Mitchell and Zymanski.32 At low concentrations, absorbance increases linearly with increase in concentration. This relationship provides linear plots with a slope of 1.0 when the logarithm of absorbance is plotted versus the logarithm of concentration.As the concentration increases, the absorption at the peak center reaches a maximum determined by the stray light. At this point, the calibration curves for LS-AAS will reach a plateau. With a continuum source, however, absorption in the wings of the profile can be measured. At any wavelength in the wings, Fig. 5 Vertical resolution of absorbance corresponding to absorbance absorbance will increase linearly with increase in concenas a function of height in the furnace.Pixel 5 is approximately the tration. However, the wings broaden as a function of the bottom of the furnace and pixel 25 is the top of the furnace. square root of the concentration, so the wavelength integrated Absorbance was measured over 200 ms intervals centered at (#) 2.2, absorbance will also increase with the square root of the (%) 2.4 and (6) 2.6 s (the peak maximum) for 500 pg of Cu concentration. This relationship at high concentrations pro- (324.7 nm).vides a linear plot with a slope of 0.5 on a log–log plot, as shown in Fig. 6. it is only 12%. This lower percentage may be due to the larger The inflection point, the point where the calibration curve mass of Cu used in Fig. 5 (500 pg) as opposed to that of makes the transition from a slope of 1.0 to a slope of 0.5, is McNally and Holcombe (200 pg). determined by the ‘a-value’ and the hyperfine splitting of the The data in Fig. 5 demonstrate the suitability of the absorption profile.The a-value is the ratio of the collisional two-dimensional CCD array for measuring intensities and width to the Doppler width and determines the width of the computing absorbance as a function of height in the furnace. absorption profile. The a-value will vary between elements and With the detector described in Table 2, it is now possible to for each element as a function of temperature. Hyperfine measure routinely the inhomogeneity of the analyte distrisplitting is determined by the coupling of the electron trans- bution in the furnace.Reading the full array (80×80 pixels), itions and determines the number of components of the however, requires a read rate 80 times faster than that required absorption profile. Elements with low a-values and fewer if the columns were binned or if a linear array of 80 pixels components will have deeper, narrower profiles and elements with an 8051 aspect were used. Further studies will determine with a larger a-values and more components will have shal- whether the increased linearity of the analytical data justifies lower, broader profiles.The former elements will reach the the increased demand for data acquisition speed. stray light limit sooner and have inflection points at lower concentrations. Calibration curves In Fig. 6, absorbance is plotted versus the normalized concentration, i.e., the concentration of the standards divided Fig. 6 shows calibration curves for Ag (328.1 nm), Cd by the intrinsic concentration (the concentration necessary to (228.8 nm) and Pb (283.3 nm)31 that consist of two linear give an integrated absorbance of 0.0044). At lower concen- regions and cover 5–6 orders of magnitude of concentration. trations, each normalized concentration will have the same A major detraction for LS-AAS has always been the relatively absorbance even though the shapes of the absorption profiles short linear region of the calibration curves, from 2.5 to 3 are diVerent.In Fig. 6, the inflection point for Cd occurs at a orders of magnitude of concentration. The limits for the linear lower concentration than that of Ag and Pb. This indicates that the resultant absorption profile for Cd is narrower and deeper than those of Ag or Pb. The calibration points for all three elements in Fig. 6 can be fitted with a single calibration curve shape.33 Half of a hyperbola was rotated and the parameters were adjusted to provide a curve with a slopes of 1.0 and 0.5 in the linear regions. This shape was fit to all three data sets (solid lines in Fig. 6) by simply oVsetting the x and y axis coordinates. Two calibration standards (one in each linear region) are suYcient to determine a unique calibration curve, but four standards provide better reproducibility for the curves. Obtaining more information from ETAAS In light of the recent interest in metal speciation, it is ironic that researchers in the field of AAS have spent many years trying to optimize furnace parameters in order to force diVerent Fig. 6 Calibration curves for (6) Ag (328.1 nm), (%) Cd (228.8 nm) species into uniform behavior, i.e., volatilization of all metal and (1) Pb (283.3 nm). Absorbance integrated with respect to species as a single analytical peak, regardless the chemical wavelength and time is plotted versus normalized concentration (concentration divided by the intrinsic concentration).environment. It seems reasonable that use of less than J. Anal. At. Spectrom., 1999, 14, 137–146 143‘optimum’ parameters would oVer more information about the metal species. It is also true that analytical peaks acquired using ‘optimum’ parameters oVer information that is not currently being used. Alternative approaches to furnace design, temperature programs and data processing can provide much more information about the sample species and how they compare with known standards.The peak shape contains a significant amount of additional information about the analyte with respect to condensed and gas phase interactions. Researchers have known for years that the leading edge of the analytical peak can be used to determine the activation energy of the metal, an indication of the chemical bond(s) that had to be broken to allow volatilization.34 Harnly35 showed that the times for the appearance, peak maximum and decay to predetermined values could be used to predict interferences.It was later shown that the phase angles of the Fourier transform of the analytical peak were more sensitive and precise indicators of diVerences between Fig. 8 Thermal separation of heme Fe and inorganic Fe (248.3 nm) in a two-step furnace. the behavior of the standards and the samples.36 Most recently, it has been shown that principle component analysis is the best means of incorporating data from all the phase angles into a comprehensive evaluation of the signal behavior.37 are not necessarily suYcient to induce atomization. The solution to this problem is to use a two-step furnace.With this Fig. 7 shows the principal components analysis of the Fourier transform phase angles for the determination of Pb in furnace, the cup temperature can be slowly ramped to maximize thermal separation and the furnace can be held at a 5% HNO3 (Group A), in 5% HCl (Group B) and in various concentrations of NaCl in both HNO3 and HCl (Groups C temperature suYcient to atomize any of the compounds.Since the temperature of the furnace is held constant, diVusion from and D, respectively). The arrows show that there was a systematic shift in position with respect to the standard the furnace will be constant and the time integrated absorbance will be independent of the introduction rate from the cup. The concentrations in the HNO3 and with the NaCl content in both matrices. The standards in 5% HCl (Group B) showed constant diVusion rate will also allow one set of standards to be used for all compounds (peaks).no systematic shift in position with respect to concentration. The recoveries for 2 ng of Pb in 0.1, 0.2, 0.5 and 1.0% NaCl Fig. 8 shows the determination of heme Fe (Fe2+ in protoporphyrin IX) and inorganic Fe. The heme Fe content in 5% HNO3 (by mass) were 90, 80, 72 and 64%, respectively. The recoveries in similar concentrations of NaCl in 5% HCl of foods is of importance to nutritionists because it is absorbed by the body at about three times the rate of inorganic Fe.were 80, 74, 72 and 69%, respectively. Unfortunately, there are not enough repeat determinations to assess the precision Heme Fe sublimes at about 250 °C whereas inorganic Fe volatilizes around 1800 °C. It can be seen that the peaks are of the measurements. Certainly, however, the data suggest the diagnostic potential of the peak shape data. nicely resolved. The two step furnace provides an ideal means of quickly determining the two Fe concentrations.At this A second approach to acquiring more information about the analyte is to slow the temperature ramp of the furnace to stage, more work is needed to determine the best method of lysing the sample matrix so that the heme Fe is not trapped allow the detection of species with diVerent volatilities. Unfortunately, temperatures suYcient to volatilize the analyte and thermally degraded before it can sublime.The temporal resolution capabilities of the two-step furnace are appropriate for metals that have high volatilities or will not thermally degrade. Several applications suggest themselves, most notably the determination of elemental species of environmental interest such as the various organic forms of As, Hg, Sn and Tl. Future of multi-element AAS The instrument described in Table 2 and throughout this paper has unique analytical capabilities. As a single-element instrument, it surpasses LS-AAS with respect to absorbance accuracy, detection limits, calibration, spectral and spatial information and the portability of the figures of merit.The general analytical features of this instrument, when used with a graphite furnace atomizer, are summarized in Table 5. Such an instrument, at the listed cost, would be competitive in the market place. The real advantage of CS-AAS, however, is multi-element determinations. Fig. 7 Principal component analysis of phase angle arrays for varying Construction of a multi-element instrument, incorporating masses of Pb (283.3 nm) in matrices of 5% HNO3, 5% HCl and various concentrations of NaCl.(A) 0.2–4.0 ng of Pb in 5% HNO3; the characteristics listed in Table 2, would require the developdirection of arrow indicates direction of increasing Pb mass. (B) ment of a new detector. Although any one of several existing 0.2–4.0 ng of Pb in 5% HCl; there was no correlation between the e� chelle spectrometers would be suitable, a solid state detector principal component scores and the Pb mass.(C) 2.0 ng of Pb in 5% would have to be developed consisting of approximately 45 HNO3 and 0.1–1.0% NaCl (by mass); arrow indicates the direction two-dimensional CCDs. In theory, one possibility is a mono- of increasing NaCl concentration. (D) 2.0 ng of Pb in 5% HCl and lithic detector in which the two-dimensional CCDs are embed- 0.1–1.0% NaCl (by mass); arrow indicates the direction of increasing NaCl concentration.ded. Such a detector would be analogous to the segmented 144 J. Anal. At. Spectrom., 1999, 14, 137–146Table 5 Summary of characteristics of single-element CS-AAS instruments as to how this principle is applied. These diVerdescribed in this paper ences may be advantageous or disadvantageous for CS-AAS. The flexibility of the Iris and Vista in selecting analytical Feature Evaluation wavelengths is a distinct advantage for CS-AAS.The fixed wavelength position of the arrays of the Optima results in Detection limits 0.1–10 pg Calibration range 105–106 omission of about one third of the resonance wavelengths routinely used for AAS. With the Iris and Vista, it is only Interferences— necessary to program data acquisition for a diVerent suite of Stray light Severe—corrected with oV-order measurements Chemical Moderate—correctability depends on furnace sub-arrays. design The low pixel read noise of the Optima and Vista is preferred Spectral—broad Severe—corrected with oV-line measurements over the higher levels of the Iris.Whereas multiple non- Spectral—line Rare—correctability varies with overlap destructive reads of the CID can eVectively reduce the read severity noise for AES applications, the rapid, transient signals of the Mass tolerance >100% m/m Resolution 2 pm at 200 nm furnace will not permit suYcient multiple reads with CS-AAS. Cost ~$35 000 Consequently, the eVective read noise of the Iris will always be greater than that of the Optima or Vista.Of course, the read noise of the CID is still suYciently low that only a few elements in the UV would be rendered read noise limited. CCD detector described by Barnard et al.,38 except that each The Iris is the only instrument with the inherent ability to detector would have two dimensions. Since far fewer wavemeasure intensities between orders for correction for say lengths are used for AAS (compared with AES), only one light.In the single-element mode, a double monochromator array would be necessary for each element routinely run by can be used to reduce the stray light arising from the CS. In ETAAS, approximately 45. Such a detector, however, would the multi-element mode of operation, this is not possible and be expensive. the far stray light (from flaws in the grating) and stray light The most obvious means of reducing the complexity of the from overlapping adjacent orders can be very high.In the detector is to sacrifice the vertical dimension for height resomulti- element mode, the Iris is the only instrument that oVers lution of the furnace. If this is done, then a series of linear stray light correction capabilities. arrays can be used. The next logical question is whether any The high quantum eYciency of the Optima and Vista arises of the solid state detector–e� chelle spectrometer instruments from the linear nature of the CCD arrays.Without overlying that have been developed for AES (Table 6) are suitable for control lines, quantum eYciencies as high as 50% at 200 nm multi-element CS-AAS. While such instruments would not be can be achieved without the use of a fluorescent coating. The ideal for AAS, the lack of development costs is appealing. two-dimensional nature of the CID requires a grid of control Table 6 presents the characteristics of the three major e�chelle lines and reduces the quantum eYciency to about 34% using spectrometer–array detector instruments which have been a fluorescent coating.developed for ICP-AES.38–41 Each incorporates the same The high luminosity (Ll) of the Optima predicts better fundamental principle; a sub-array of pixels is used for each detection limits than the other two instruments. Higher lumin- wavelength of interest. This principle is also valid for CS-AAS. There is, however, a great deal of diVerence between the osity means greater intensity striking each pixel and, from Table 6 Comparison of commercially available atomic emission array detectors Optimaa Irisb Vistac Spectrometer characteristics— Type E� chelle E� chelle E�chelle Focal length/mm 504 384 400 Blaze angle (°) 63.4 19.5 44.8 Grating area/mm×mm 80×160 40×40 50×70 RLDd (200 nm)/nm mm-1 0.100 0.735 0.252 Entrance slit height/mm 250 58 50 Entrance slit width/mm: Normal 62 58 25 High resolution 31 32 – Spectral bandwidth/pm: Normal 6.2 43 6.3 High resolution 3.1 23 – Lle (200 nm)/mm2 nm-1 0.028 0.00038 0.0015 Detector characteristics— Type of array CCD CID CCD Segmented-linear Two dimensional Diagonal-linear Number of arrays 224 1 140 Array size(s) 1×20 to 1×80 512×512 1×750 to 1×1521 Number of pixels 6336 262 144 70 908 Pixel size/mm 12.5×170 (UV) to 12.5×80 (VIS) 28×28 12.5×105 (UV) to 25×45 (VIS) Read noise (e-) 15 260 <10 Quantum eYciency 56 34 50 (200 nm) (%) Sub-array processing Inherent Yes Yes Binning capability Not necessary Yes Yes Wavelength coverage/nm 167–782 170–800 165–785 aPerkin-Elmer.38,39 bThermo Jarrell Ash (Franklin, MA, USA).40 cVarian, Optical Spectroscopy Instruments (Mulgrave, Australia).41 d Reciprocal linear dispersion.eLuminosity per unit wavelength as defined in ref. 39. J. Anal. At. Spectrom., 1999, 14, 137–146 145P. L. Lundberg and B. Radziuk, J. Anal. At. Spectrom., 1997, eqn. (3), less absorbance noise. Since the wavelength and time 12, 617.integrated absorbance is constant, lower absorbance noise 14 J. M. Harnly, R. E. Fields and M. Schuetz, J. Anal. At. Spectrom., specifies better S/Ns and lower detection limits if the same in the press. furnace is used. 15 R. L. Cochran and G. M. Hieftje, Anal. Chem., 1977, 49, 2040. The Optima and Vista oVer seven times better resolution 16 R. E. Sturgeon and C. L. Chakrabarti, Prog. Anal. At. Spectrosc., 1978, 1, 5. than the Iris in the ‘normal’ mode of operation.The resolution 17 B. L’vov, Spectrochim. Acta, Part B, 1978, 33, 153. of both the Optima and the Iris can be improved by a 18 A. Kh. Gilmutdinov and J. M. Harnly, Spectrochim. Acta, Part B, factor of approximately two in the ‘high’ resolution mode. As 1998, 50, 1003. discussed earlier, the S/N is independent of the spectral 19 Spectraspan IIIB Emission Spectrometer Operators Manual, Part bandwidth. The bandwidth is an important factor, however, Number 1506029, Spectrametrics, 1977.in resolving spectral line interferences. 20 Specifications for FastOneTM, PixelVision, Beaverton, OR, 1997. 21 W. Slavin, Graphite Furnace: a Source Book, Part Number Although no specifications are presented in Table 6 for the 0993–8139, Perkin-Elmer, Norwalk, CT, 1984. speeds of operation, the general characteristics of each instru- 22 C. M. M. Smith and J. M. Harnly, Spectrochim. Acta, Part B, ment allow the determination of at least 10 sub-arrays every 1994, 49, 387. 17 ms (60 Hz). The number of elements determined may 23 J. M. Harnly, Spectrochim. Acta, Part B, 1993, 48, 909. require compromises with respect to integration times, shift 24 W. Frech and S. Jonsson, Spectrochim. Acta, Part B, 1982, 37, rates and frequency of analog-to-digital converter operation 1021. 25 SIMAA 6000 Atomic Absorption Spectrometer, Part Number which will alter the S/N. Exact values for detection limits are B050–6158, Perkin-Elmer, Norwalk, CT, 1989. diYcult to predict. 26 M. Schuetz, PhD Thesis, Technical University of Berlin, 1997. It can be seen that each of the commercially available 27 NIST Circular 488, Section 4, Ultraviolet Multiplet Table (Z=1 e� chelle–solid state detector instruments oVers diVerent capa- to 64), and Section 5, Ultraviolet Multiplet Table (Z=72 to 88), bilities. None is ideal. The cost of development and construc- National Institute of Standards and Technology, Gaithersburg, tion of the ‘ideal’ detector would be expensive. In the current MD, 1978. 28 J. M. Harnly, unreported results, 1998. economic climate, where development of technology for future 29 A. Kh. Gilmutdinov, B. Radziuk, M. Sperling and B. Welz, products is sacrificed for quarterly profits, the cost of the Spectrochim. Acta, Part B, 1996, 51, 1023. detector may be suYcient to block the development of multi- 30 J. McNally and J. A. Holcombe, Anal. Chem., 1987, 59, 1105. element CS-AAS, regardless of the enhanced analytical 31 D. N. Wichems, R. E. Fields and J. M. Harnly, J. Anal. At. capabilities. Spectrom., 1998, 11, 1277. 32 A. C. G. Mitchell andM.W. Zymansky, Resonance Radiation and Excited Atoms, Cambridge University Press, Cambridge, 1961. References 33 J. M. Harnly, C. M. M. Smith and B. Radziuk, Spectrochim. Acta, Part B, 1996, 51, 1055. 1 A. Walsh, Spectrochim. Acta, Part B, 1955, 7, 108. 34 R. E. Sturgeon, C. L. Chakrabarti and C. H. Langford, Anal. 2 C. T. J. Alkemade and J. M. W. Milatz, J. Opt. Soc. Am., 1955, Chem., 1976, 48, 1792. 45, 583. 35 J. M. Harnly, J. Anal. At. Spectrom., 1988, 3, 43. 3 V. A. Fassel, V. G. Mossotti, W. E. Grossman and R. N. Knisely, 36 J. M. Harnly, J. Anal. At. Spectrom., 1988, 3, 485. Spectrochim. Acta, Part B, 1966, 22, 347. 37 J. M. Harnly, Paper presented at the Pittsburgh Conference and 4 McGee and J. D. Winefordner, Anal. Chem., 1967, 37, 429. Exhibition, New York, March 5–9, 1990, Paper 572. 5 W. Snelleman, Spectrochim. Acta, Part B, 1968, 23, 403. 38 T. W. Barnard, M. J. Crockett, J. C. Ivaldi, P. L. Lundberg, 6 R. C. Elser and J. D. Winefordner, Anal. Chem., 1972, 44, 698. D. A. Yates, P. A. Levine and D. J. Sauer, Anal. Chem., 1993, 7 J. G. Nitis, V. Svoboda and J. D. Winefordner, Spectrochim. Acta, 65, 1231. Part B, 1972, 27, 345. 39 T. W. Barnard, M. J. Crockett, J. C. Ivaldi and P. L. Lundberg, 8 C. Veillon and P. Merchant, Appl. Spectrosc., 1973, 27, 361. Anal. Chem., 1993, 65, 1225. 9 P. N. Keliher and C. C. Wohlers, Anal. Chem., 1974, 46, 682. 40 M. J. Pilon, M. B. Denton, R. G. Schleicher, P. M. Moran and 10 A. T. Zander, T. C. O’Haver and P. Keliher, Anal. Chem., 1976, S. B. Smith, Jr., Appl. Spectrosc., 1990, 44, 1613. 48arian Optical Spectroscopy Instruments, Sugarland, TX, 11 J. M. Harnly, T. C. O’Haver, B. Golden and W. R. Wolf, Anal. personal communication, 1998. Chem., 1979, 51, 2007. 12 J. M. Harnly, J. Anal. At. Spectrom., 1993, 8, 317. 13 J. M. Harnly, C. M. M. Smith, D. N. Wichems, J. C. Ivaldi, Paper 8/07586G 146 J. Anal. At. Spectrom., 1999, 14, 137–146
ISSN:0267-9477
DOI:10.1039/a807586g
出版商:RSC
年代:1999
数据来源: RSC
|
8. |
Hyphenated vapour generation atomic absorption spectrometric techniques |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 147-162
Dimiter L. Tsalev,
Preview
|
|
摘要:
Hyphenated vapour generation atomic absorption spectrometric techniques† Invited Lecture Dimiter L. Tsalev Faculty of Chemistry, University of Sofia, 1 James Bourchier Blvd., Sofia 1126, Bulgaria Received 18th September 1998, Accepted 5th November 1998 1 Aims and scope 100 articles and has authored and co-authored six monographs, 2 Modern flow techniques in VGAAS (FI-VGAAS chapters and reviews on AAS in occupational and environmental and CF-VGAAS) health practice, hydride generation AAS, flow injection sample 3 On-line decompositions (FI-MWD-VGAAS, CF- treatment and enrichment by vapour generation and extraction, MWD-VGAAS, FI-UV-VGAAS, etc.) and chemical modification in electrothermal AAS.Dr. Tsalev 4 On-line pre-reduction was Programme Chairman of the XXVI CSI and organiser of 5 On-line preconcentration and separation (FI-VG- the Post-symposium on Electrothermal Atomisation in 1989 and ETAAS, FI-ion-exchange-VGAAS, HG-CT-AAS, has been on the Editorial Boards of Spectroscopy Letters and etc.) Spectrochimica Acta Part B. 5.1 VG-ETAAS 6 Hyphenated techniques for speciation analysis (VG-CT/GC-AAS, HPLC-VGAAS, HPLC-UVChemical vapour generation (VG) atomic absorption VGAAS, HPLC-MWD-VGAAS, etc.) spectrometry (VGAAS) such as the cold vapour (CV) tech- 6.1 On-line speciation by non-chromatographic nique for mercury, hydride generation (HG) for antimony, HGAAS techniques arsenic, bismuth, germanium, lead, selenium, tellurium and tin 6.2 Coupling with GC and in situ ethylation for mercury, lead, selenium and tin to 6.3 Coupling with HPLC their volatile derivatives has become popular for the determi- 6.3.1 Scope, sensitivity and LODs nation and speciation of these diYcult analytes and their 6.3.2 Hardware complexity organoelement species, particularly in the important appli- 6.3.3 Mercury speciation cation field of biological and environmental monitoring. In 6.3.4 Availability and validation this overview are highlighted the current state, analytical scope, 7 Conclusion 8 References performance characteristics, advantages, limitations and prospects of hyphenated VGAAS techniques on the basis of the author’s own research supplemented with selected examples from the literature (256 references).The term ‘hyphenated’ is employed in its common grammatical sense, thus comprising not only the coupled chromatographic techniques for speciation analysis, such as high-performance liquid chromatography (HPLC)-VGAAS, VG-gas chromatography (GC)-AAS, VG-cryogenic trapping (CT)-GC-AAS, etc., but also a variety of modern, automated flow injection (FI )- and continuous flow (CF)-VGAAS approaches utilising eYcient hyphenations with on-line sample/analyte treatments (pre- or post-VG or both), such as sample introduction, decomposition, prereduction, enrichment, non-chromatographic separations, etc.Particularly highly ranked are the modern FI-VGAAS techniques with sensitive atomisation modes, such as quartz tube atomiser (QTA) cells, ‘flame-in-tube’ (FIT) atomisers and in-atomiser trapping of chemical vapours with subsequent electrothermal AAS (ETAAS) quantification (FI-VGDimiter Tsalev obtained his Ph.D.in Analytical Chemistry from ETAAS), as well as eYcient sample pre-treatment means such Moscow State University, Russia in 1972 and Doctor of Sciences as on-line ultraviolet photooxidation (UV) or microwave in Chemistry from the Bulgarian Academy of Sciences in 1996.decomposition (MWD) of diVerent organoelement species He is Associate Professor of Analytical Chemistry at the which better suit the requirements for automation, scaling University of Sofia. Professor Tsalev has lectured in environmen- down, adequate sensitivity, interference control, sample tal and advanced analytical chemistry, instrumental techniques, throughput rates, reliability, commercialisation of hardware, chemical control for biotechnologists and has given invited budget and routine applicability, viz., FI-MWD-CVAAS for lectures in the field of atomic absorption spectrometry (AAS) Hg with or without amalgamation; FI-MWD-HGAAS for As, at scientific conferences and seminars.He has published over Bi, Pb, Se and Sn; FI-UV-HGAAS for As; VG-ETAAS for As, Bi, Cd, Hg, Ge, Pb, Sb, Se, Sn and Te; FI-UV-HGETAAS for As and Se; HPLC-VGAAS (QTA) for As, Hg †Presented at The Ninth Biennial National Atomic Spectroscopy Symposium (BNASS), Bath, UK, July 8–10, 1998.and Sb; HPLC-UV-VGAAS (QTA) for As, Hg and Sn; HPLCJ. Anal. At. Spectrom., 1999, 14, 147–162 147MWD-HGAAS for As and Se; VG-CT/GC-AAS (FIT) for As, Hg, Pb, Sb, Se and Sn; FI-HGAAS or FI-HG-ETAAS for redox or ‘first-order’ speciation such as ‘toxicologically relevant arsenic’, arsenite, selenite and tellurite and for on-line preconcentration and separation. Among the chemical systems utilised for on-line UV photooxidation or MWD are mostly recommended the alkaline (for Hg and Se) or acidic (for Bi, Hg, Pb and Sn) peroxodisulfate and acidic bromination reagents (bromide–bromate) for Bi, Hg and Se.On-line prereduction to lower, reactive oxidation states is expedited by hot L-cysteine for As, monomethylarsonate and dimethylarsinate and by hot bromination reagents or hot concentrated hydrochloric acid for Se. An important niche of the coste Vective hyphenated VGAAS techniques with prospects for their establishment in routine practice for automated trace Fig. 1 Cumulative number of publications on some hyphenated element determinations and speciation analysis versus other VGAAS techniques. emerging and more sensitive techniques is defined. confined mainly to papers from the last decade. Reference is The term ‘hyphenated techniques’ in vapour generation atomic made in appropriate places to the relevant recent books and absorption spectrometry (VGAAS)1 is employed in this text monographs1–3,6–9,35,36 and chapters and reviews on in its common grammatical sense, thus comprising not only VG,37 HG,1,37–39 ethylation,40 VG-ETAAS,19 MWD,41,42 UV the coupled chromatographic techniques for speciation photooxidation43,44 and speciation.1–6 analysis1–6 but also a variety of VGAAS methods with on-line sample/analyte treatments (pre- or post-VG or both), such 2 Modern flow techniques in VGAAS (FI-VGAAS as sample introduction, decomposition, pre-reduction, enrichment, non-chromatographic separations, etc.and CF-VGAAS) Methodologically, hyphenation could be considered as a gen- Various chemical systems for VG in analytical atomic eral trend and a historic stage during the process of developspectrometry for up to 30 elements have been documented ment of most modern instrumental techniques, aimed at (Table 1)1 but only a few of these systems with less than one- improving their performance characteristics, versatility and third of the analytes exhibit adequate, competitive performance automation, e.g., GC-MS (gas chromatography-mass speccharacteristics.Most attractive are the chemical systems based trometry), ICP-MS (inductively coupled plasma mass specon (i) reactions in aqueous media, (ii) with fast kinetics of trometry), ETV-ICP-AES (electrothermal vaporization VG and easy stripping oV of vapours from solution, (iii) inductively coupled plasma atomic emission spectrometry), reliable interference control, (iv) with high and reproducible high-performance liquid chromatography (HPLC)-ICP-MS, chemical yields of volatile products which are (v) readily etc.It is worth noting that the most sensitive AAS techniques, atomised in quartz tube atomiser (QTA) cells or ‘flame-in- cold vapour AAS (CVAAS), hydride generation AAS tube’ (FIT) atomisers rather than in flames, so as to provide (HGAAS) and electrothermal AAS (ETAAS), had also been adequate sensitivity and low limits of detection (LOD). historically spelled with hyphenation (CV-AAS, HG-AAS and While the above requirements are already stringent enough, ET-AAS) for almost two decades, until they gained maturity a few more could be added (blank control, economic consider- and reached a stage of interfacing and integration within AAS ations, preference for room temperature reactions, operational apparatus to lose their old hyphens and eventually to acquire hazards, etc.), thus limiting the current practical choice to new, furthermore eYcient hyphenations with flow injection three VGAAS systems: (i) HG with sodium tetrahydroborate, (FI)-HGAAS,1,7–9 on-line microwave decomposition (MWD) NaBH4, or (rarer) KBH4 for As, Bi, Sb, Se, Sn and Te; (ii) or on-line UV photooxidation (UV), viz., FI-MWDCV generation for Hg with NaBH4 or SnII reductants; and CVAAS,10–13 FI-MWD-HGAAS,10,12,13 continuous flow (iii) generation of ethyl derivatives (ethylation) with sodium (CF)-UV-CVAAS,14,15 FI-UV-HGAAS,16–18 vapour genertetraethylborate (STEB, NaBEt4)40,45 for Hg, Pb, Se and Sn.ation with in-atomiser trapping and ETAAS quantification In situ ethylation-HGAAS is a promising (yet not fully (VG-ETAAS,19 HG-ETAAS,20–27 CV-ETAAS28–30), explored) alternative to common HGAAS and CVAAS tech- VG-GC-AAS,31–33 HPLC-UV-HGAAS,16,34 etc. niques in some diYcult analytical tasks (for a review, see 1 Aims and scope Table 1 Chemical vapour generation in analytical atomic spectrometry The aim of this overview is to highlight the current state, Chemical system for analytical scope, performance characteristics, advantages, limi- vapour generation Analytea tations and prospects of hyphenated AAS techniques with chemical vapour generation.Emphasis is placed on the Hydride generation (HG) As, Bi, Ge, (In), Pb, Sb, Se, Sn, Te, (Tl) Cold vapour (CV) Cd, Hg, (Cu?) author’s own research in this field which is supplemented by technique selected examples from the literature.Modern approaches Ethylation Bi, Cd, Co, Ge, Hg, Pb, Se, Sn, Tl based on FI and CF performance which better suit require- Butylation Be, Ga, Hg, Pb, Sn, Zn ments for automation, scaling down, sample throughput rates, Carbonyl generation Co, Fe, Ni on-line treatments, reliability and routine applicability are Chlorides As, Bi, Cd, Ge, Mo, Pb, Sn, Tl, Zn particularly favoured. Related papers on electrothermal Fluorides Ge, Mo, Re, U, V, W Dithiocarbamates Co, Cr, Cu vapour generation, oV-line decomposition, pre-reduction and b-Diketonates Al, Co, Cr, Cu, Fe, Mn, Ni, Pb, Zn enrichment, collecting chromatographic eZuent fractions, Miscellaneous VG B(CH3O)3, OsO4 coupled chromatography–graphite atomiser AAS and nontechniques AAS detectors are not considered.Since the number of relevant aSpecies in bold italic indicate straightforward analytical performance; papers on hyphenated VGAAS well exceeds 400 and substanspecies in parentheses indicate serious problems with that particular tial activity and progress have been registered in recent years analyte.(see Fig. 1 for trends in publications), the list of citations is 148 J. Anal. At. Spectrom., 1999, 14, 147–162Table 2 Advantages and limitations of chemical vapour generation in AAS Advantages Disadvantages and limitations High chemical yields and transport eYciency Complex, on-line chemistry involved Separation from troublesome matrices Specific sample pre-treatment required Enrichment EVect of oxidation and binding state of analyte Automation (in FI and CF mode) Chemical interferences Good sample throughput rates (e.g., 40–60 h-1) Foaming and aerosol formation Competitive LODs (0.02–0.3 ng or 0.1–0.3 mg l-1) and RSDs (1–3%) Numerous instrumental and chemical parameters for optimisation Commercially available accessories Applicability to a limited number of analytes Economically aVordable Limited linear range (<103) Speciation potentialities Poor multi-element capabilities Rapsomanikis40): first for derivatization of the ionic alkylated species of Hg,46,47 Pb,45,48 Se49 and Sn50,51 in speciation analysis by cryogenic trapping (CT)-GC-AAS,45–47,50 GCAAS51 or HPLC-ethylation-AAS48 and also in determinations of the total amount of some hydride-forming elements (HFEs) which are ‘problematic’ in HGAAS, such as Pb (blank problems, oxidant required, low yields of HG reaction, narrow pH interval ), Sn (reagent blanks, interferences, narrow pH interval for HG) or Cd and Tl (very unstable hydrides, hardly determined by HGAAS1).Some adverse comments on ethylation could be added: expensive reagent, high toxicity of ethyl derivatives and resorting to FIT atomisation. In Table 2 are summarised and approximately ranked in Fig. 2 Schematic representation of possibilities for on-line decomposition and pre-reduction in VGAAS, with analytes shown on the right. their decreasing order of importance the advantages and limitations of VGAAS, considering that FI or CF operation modes rather than batch type of VG are used.A detailed of residual oxidants, reagent blanks, etc., as discussed in detail comparative treatment of these operational approaches has for individual analytes and matrices in relevant texts.1,9,35,36 been given elsewhere.1,39 The superiority of FI-VGAAS Some possibilities for on-line decomposition in flow systems hyphenation in modern AAS technology is nowadays widely are shown schematically in Fig. 2; generally they fall into two recognised, both in view of direct automated analyses with categories: (i) either involving wet chemical attack prior to high sampling frequencies (60 h-1), remarkable precision VG with strong oxidants at room or elevated temperature (within-run RSDs 1–2%), better interference control, low (microwave- or UV-assisted) or otherwise (ii) relying on highabsolute LODs (0.01–0.05 ng), etc., and moreover in its further temperature pyrolysis (700–800 °C)76,77 of GC-separated hyphenations aimed at on-line decompositions (e.g., FI-MWD- organomercury derivatives46,75–78 or on ‘thermochemical VGAAS,10–13,52–62 FI-UV-VGAAS16,17,63–65), on-line hydride generation’ (THG)79 in a reducing H2–O2 ‘flame-inenrichment (e.g., FI-HG-ETAAS,1,19,66 FI-ion-exchange- tube’ of thermosprayed (TS) HPLC eZuents.48,79–89 While all VGAAS,67–71 on-line matrix isolation by ion exchange,72–74 thermochemical atomisation means (small pyrolysis chambers, etc.) and on-line speciation methods.The FI instrumentation heated QTA and FIT) ensure eYcient atomisation for organo Vers the most versatile performance, providing computer omercury derivatives such as CH3HgH46,78 and Et-, Et2-, and controlled options for volume- and time-based sampling (e.g., phenyl (Ph)-mercurials,75 the TS–THG interface for HPLC- 0.05–1 and 2–10 ml in FI and CF modes, respectively), AAS speciation analyses for As,79,81,83–85 Cd,88 Hg,88 multiple injections in FI-CVAAS with amalgamation Pb,48,80,81,88,89 Se79,81,82,84–87 and Sn81 involves large flows of (2–10 ml ), merging zones, synchronisation with other events, FIT support gases (H2 and O2 at critical ratios and total flow etc.Batch type VG is more diYcult to automate and more rates up to 1–2 l min-1) and lacks sensitivity (LODs 1 ng) susceptible to liquid phase interferences; hence the use of batch and operational safety in order to be implemented in generators is steadily declining, being confined to some precon- hyphenated instrumentation.centration procedures from large sample volumes (5–50 ml ) On-line wet chemical oxidations in FI- or CF-VGAAS such as cryogenic trapping or VG-ETAAS with in-atomiser systems have been studied intensively in recent years and trapping—with excellent LODs down to 1 ng l-1 levels, brought to commercially available accessories. So far only a though.1 few oxidation reagents have shown promising characteristics and are compiled and graded in performance in Table 3; some selected applications to real analytical matrices are briefly 3 On-line decompositions (FI-MWD-VGAAS, outlined in Table 4.The most important observations and CF-MWD-VGAAS, FI-UV-VGAAS, etc.) conclusions concerning on-line decompositions in flow systems are itemized below. Most VGAAS procedures for the determination of total element contents rely on elaborate and time-consuming sample (a) Long reaction times (>1.5–2 min) are impracticable because of excessive dispersion of sample zone, impaired pre-treatment step(s), typically performed oV-line in the batch mode, which is one of the main drawbacks of this technique sample throughput rates and back-pressure of long reactor coils.For example, knitted reactors (KR) with id 0.5 mm and (see Table 2). Pre-treatment steps are aimed at decomposition of the sample matrix, digestion of organoelement species of length 15 m already entail problems.(b) Only a few chemical oxidants provide adequate reaction the analyte and eventually pre-reduction of certain HFEs to their reactive species (SeIV, TeIV) or more sensitive oxidation rates, e.g., alkaline peroxodisulfate, acidic bromination, acidic permanganate and (in some cases) acidic peroxodisulfate (see states (AsIII, SbIII). Sample treatment is the major source of analytical errors due to volatilization and retention losses, Table 3).Handling KMnO4 solutions is inconvenient because of strong adsorption on plastic tubing and other surfaces with incomplete oxidation and/or pre-reduction, interference eVects J. Anal. At. Spectrom., 1999, 14, 147–162 149Table 3 Selected reagents for on-line wet chemical decomposition in VGAASa Room Thermal Microwave UV UV+thermal Analytes and Reagent temperatureb heatingb heatingb irradiationb heatingb selected references K2S2O8–NaOH – +(As) ++(As, Se) +++(As) ? As,13,16,18,57,64,90–95 Se96 K2S2O8–acid(s) +(Hg) + ++ ++ +++ As,71,97 Bi,10–13 Hg,17,98 Pb,13 Sn13,34,63,65 K2S2O8–KMnO4–acid(s) +(Hg) +(Hg) ++ + ? Hg57,62 BrO3-–Br-–acid(s) +(Hg) + +++ +(Sn) +(Sn) Bi,10,12,13 Hg,10–13 Se,99–102 Sn12,13,63,65 BrO3-–Br-–K2S2O8–acid + +++ As,52,53 Hg53,103 BrO3-–Br-–KMnO4–acid +(Hg) + +++ Hg54–57 K2Cr2O7–HNO3 with or +(Hg) +(Hg) Hg15,104–106 without Cd2+ catalyst KMnO4–HNO3–HCl ++(Hg) Hg62 aFor selected applications to real matrices, see also Table 4.bPerformance ranked as not applicable at all (–), problematic or lack of data (?), good (+), straightforward (++) and a technique of choice (+++). build-up of hydrated MnIV oxide products. Therefore, per- (e) Dispersion can be reduced by a factor of 2–6 by employing ‘knotted’ or ‘knitted’ PTFE tubing reactors54 and manganate is always added on-line,54–57,62 and adequate cleaning of the FI system is applied daily55 or more often.62 by flow segmentation with small bubbles of air34 or Ar.90 (f ) Despite the lack of comparative studies between the (c) Most of the known catalysts of wet oxidation cannot be employed because of their depressive eVects on HG (AgI, CdII, three modes of sample pre-treatment, MWD, UV photooxidation and thermostated bath (TB) heating, an attempt is made CuII) or poor eYciency (H2O2–FeII,13 OsVIII,13 F- 52,53).(d) Organic matter of biological fluids and the matrix of in Table 3 to grade these approaches on the basis of the author’s experience.10–13,63–65 From the viewpoint of hardware slurried, solubilized or extracted solid samples cannot be completely decomposed; thus most ‘decompositions’ result in sophistication, price, operational hazards and accompanying problems, the approximate ranking appears to be fact in extracting the analyte from the matrix,17,62 maintaining frothing at tolerable levels13 and digesting (isoforming) organ- MWD>UV+heating&UV>TB, and accordingly the eYciency in digesting organoelement compounds follows a oelement compounds of the analyte element—most importantly those of them which do not react with tetrahydroborate similar order: MWD#UV+heating#UV&TB.(g) Some problems and remedies in FI- or CF-MWD- reductant such as arsenobetaine (AB), arsenocholine (AC), tetramethylarsonium (Me4As+), trimethylselenonium VGAAS are worth mentioning. Sample throughput rates are generally a factor of two or more lower than those in direct (Me3Se+) and the ‘heavy’ organotins (Bu2-, Bu3-, Pr3-, Ph3-, etc.).63,65 Some monoalkylated species such as monomethylar- FI-VGAAS, viz., 20–40 h-1 without amalgamation and down to 7–20 h-1 in the amalgamation mode.Unless a pre-reduction sonate (MMA)43,65,107 and monobutyltin (BuSn3+)63,65 appear to be oxidised more slowly; likewise, Me3Se+ is step is added, a sensitivity loss by a factor of 2–8 due to measuring the less sensitive oxidised species, AsV and SbV, is extremely diYcult to decompose on-line.108 Organomercurials exhibit the best performance in both FI-MWD- entailed; some improvement can be obtained after increasing the NaBH4 concentration or the length of the reaction coil in CVAAS10–13,54–57 and FI- or CF-UV-CVAAS,14,15,109–114 and some matrices can be processed even at room temperature and the HG manifold, or both.1 Because of the inhomogeneity of power distribution within the microwave cavity and a high with higher sampling frequencies (80103 or 100 h-155), e.g., saliva,103 urine55 or HPLC eZuents.110–114 percentage of non-absorbed power, three-dimensionally disori- Table 4 Selected applications of automated flow injection on-line decomposition in VGAAS with real samples Analyte Sample Brief outline of procedurea Ref.As Urine FI-UV-HGAAS: 15×diluted urine (100 ml injections) merged with alkaline peroxodisulfate (4% m/v 63, 65 K2S2O8–1 mol l-1 NaOH); on-line UV photooxidation in a KR (tirr#90 s; 8 W), then acidified with 4 mol l-1 HCl; LOD 6.6 mg l-1; 12 h-1 As Urine FI-MWD-HGAAS with 5% m/v K2S2O8–2 mol l-1 HNO3–2 mol l-1 H2SO4 (tirr 30 s; 30W); 4 h-1; 71 part of speciation scheme Bi Urine FI-MWD-HGAAS: 1+2 dilution with a bromination reagent (2.7 mmol l-1 KBrO3–13.4 mmol l-1 10, 13 KBr–2 mol l-1 HCl); 10 m KR; 60 W; LOD 0.3 mg l-1; 30 h-1 Hg Blood FI-MWD-CVAAS: 10×diluted blood in 0.12% m/v Triton X-100–0.32% m/v KBr–0.09% KBrO3; 54, 56, 500 ml injections; on-line addition of aqueous 0.05% m/v KMnO4; 10 m KR; 30 W; LOD 0.1 and 57 1 mg l-1 with Perkin-Elmer (U� berlingen, Germany) FIMS56 and FIAS systems,54 respectively; 45 h-1 Hg Saliva FI-CVAAS: 10×diluted saliva in 3.2% m/v HCl–0.32% m/v KBr–0.09% KBrO3; 500 ml injections; 103 on-line addition of aqueous 0.25% m/v KMnO4; room temperature; LOD 0.5 mg l-1; 80 h-1 Hg Sediments FI-MWD-CVAAS: sediment slurry mixed on-line with HNO3–HCl–KMnO4 62 Hg Tissues FI-UV-CVAAS: solubilized tissue (2% m/v) in aqueous 4 g l-1 NaCl–2 g l-1 L-cysteine–90 g l-1 17 NaOH; 100 ml injections; on-line UV photolysis of CH3HgCl with 1% m/v K2S2O8–6 mol l-1 HCl; LOD 0.18 mg g-1 Hg Urine, water FI-MWD-CVAAS: urine diluted 1+5 or 1+2 (with or without amalgamation, respectively) with 10–13, 2.7 mmol l-1 KBrO3–13.4 mmol l-1 KBr–1 mol l-1 HCl–50 mg l-1 K2Cr2O7–0.1 mol l-1 HNO3; 57 LODs 0.01 and 0.2 mg l-1 with and without amalgamation, respectively Se Tap water (spiked) A ‘first-order’ FI-MWD-HGAAS speciation scheme for i-SeIV and total Se, with ‘on–oV’ operation 100 of MWD in aqueous 2.4 mmol l-1 KBrO3–2.6 mol l-1 HBr medium; LOD 0.8 mg l-1 Sn Urine FI-MWD-HGAAS: urine diluted (1+1) with 2.7 mmol l-1 KBrO3–13.4 mmol l-1 KBr–10 mmol l-1 13 HCl–1% m/v tartaric acid or with 1% m/v K2S2O8–50 mmol l-1 H2SO4–1% m/v tartaric acid; 10 m KR; 60 W; LOD 0.2 mg l-1 aAbbreviations: KR, knitted reactor; tirr, irradiation time. 150 J. Anal. At. Spectrom., 1999, 14, 147–162ented reactors (knotted reactors, KR) with an appropriate plexes),122,128–130 which play certain important roles in As geometric orientation within the cavity of the focused micro- analysis and speciation.122,128–132 The kinetics of these prewave oven are used in addition to an appropriate ballast load reduction/complexation reactions is slower for methylated (coil ) in order to avoid reflected power.10–13,52–57 The eVect of species122,128 but the MMA and DMA can still be complexed temperature on sensitivity in VGAAS is positive;10–13,96,100 within 1.5–2 min122 compared with up to 1–2 h at room however, some verse eVects of running hot samples are the temperature.evolution of water vapour and aerosols, pressure build-up and Ever since its successful introduction in 1987 by Brindle and disturbances of the liquid flow. They are treated by placing Ceccarelli Ponzoni,133 the L-cysteine reagent has been revealing downstream after the MWD oven a cooling coil54–58 and back- its valuable assets in HG methodology, which have been pressure regulator device95 or by passing the flow through a discussed in more length elsewhere by Welz and second peristaltic pump after the cooling coil.100,101 Aerosols S¡ ucmanova�,134,135 Le et al.,129 Howard and Salou128,130 and and water vapour condensates are trapped by in-line filters Tsalev et al.122,131 It is fully compatible with modern CF and from PTFE membranes10–13 or glass-fibre (for Hg only)10,11 FI systems in both oV- and on-line modes (without and with or Nafion tubing.62 An antifoam agent may be added to the heating, respectively); both pre-reduction and the subsequent NaBH4 solution, e.g., Dow Corning 110 A emulsion from HG are faster, more eYcient and tolerant to interferents than 0.04–0.08% v/v (for urine)10–13 to 0.15% v/v (with a blood with other pre-reductants and reaction media.1 At least two matrix).54,56 The composition and mass of sample and reagents extra positive assets are worth mentioning: (i) multi-element added are important for the eYciency of coupling microwave possibilities are extended to several other accompanying HFEs and UV energy to sample flow and for the completeness of (As, Bi, Ge, Sb, Sn);131 and (ii) unified HG conditions with digestion.This has been documented for various compositions improved recoveries of the total i-As (arsenite+arsenate), such as urine,58,59,63,65 urea,16 methanol,16 phosphate buVers,16 MMA and DMA are provided,128–132 with an impact on other KMnO4,55 ammonium salts16 etc.Higher concentrations of hyphenated or speciation procedures such as HG-CT/GCoxidant, longer reaction times and higher dilution factors are AAS,128–130 FI-HGAAS122,132 and HPLC-HGAAS122 in all beneficial. Applying stop–flow operation in order to water128,130 and urine analyses.129,132 improve decomposition, owing to pressurization and increased Employing pre-reductants other than L-cysteine such as reaction times, is not practical with current FI plumbing KI–HCl120,121 or KI–ascorbic acid–HCl71,116,117,119 does not because of troublesome leakages and reduced sampling show much promise for on-line pre-reduction, despite intensive frequencies.heating (95–140 °C) and high levels of reductants (15–50% m/v) and acid (2–5 mol l-1) applied. Intolerable 4 On-line pre-reduction blanks and reagent consumption, complicated manifolds71,117,119 and extra dispersion are entailed.OV-line pre-reduction (SeVI�SeIV, TeVI�TeIV, AsV�AsIII and For selenium, the best results are obtained with on-line SbV�SbIII) is an essential and well documented stage of decompositions which yield the analyte in its selenite (i-SeIV) HGAAS methodology.1,9,35,36 It is almost indispensable for state, thus radically eliminating the need of a separate pre- Se and Te, the higher oxidation states of which (selenate and reduction step; microwave or TB heating in the presence of tellurate, respectively) do not produce hydrides with tetra- KBrO3–HBr or KBrO3–KBr–HCl is typically applied.99–102,108 hydroborate, except in certain tasks such as speciation analysis Unfortunately, the Me3Se+ species have not been addressed based on selective HG from only the i-SeIV (selenite) or i-TeIV in most on-line studies (except in ref. 108), although they (tellurite) (i=inorganic), or otherwise in those favourable are present as a major component in urine samples and are cases whenever the sample decomposition method has already known to be very resistant to wet decompositions and transformed the analyte into its tetravalent state—as happens bromination.136–139 for Se in the presence of bromide.99–102 For more data on Se An alternative option of on-line pre-reduction to SeIV, and Te determination and speciation, see also the recent review although less eYcient and not as convenient, involves hydro- by D’Ulivo.115 chloric acid medium,1 the sample digests being merged In As and Sb determinations, both trivalent and pentavalent with concentrated HCl (10–12 mol l-1) on heating at forms of analytes yield hydrides but nevertheless pre-reduction 95–140 °C.69,96,123–126 Essential recent studies on the kinetics is highly desirable since the i-AsIII (arsenite) and i-SbIII (antiof bromination100,136–139 and pre-reduction with HCl1,125,140,141 monite) are determined with better sensitivity (2–8-fold). This are worth consulting for further reading.step is optional, however, and may be omitted if such a sensitivity loss can be tolerated (say in HG-ETAAS) and the analytes are already definitely transformed into their penta- 5 On-line preconcentration and separation (FI-VG- valent states, e.g., in aqua regia leachates or digests of soil, ETAAS, FI-ion-exchange-VGAAS, HG-CT-AAS, sediment or ferrous alloys.1 There are several other concerns about pre-reduction performance: slow reaction kinetics, high etc.) consumption of reagents, blanks, volatilization and retention On-line separation and enrichment in FI-VGAAS and losses, adverse eVects of residual oxidants and matrix con- CF-VGAAS have gained much popularity in recent stituents, time stability of the prepared solutions, limited years,1,7–9,36 thanks to extensive research and commercial multi-element capabilities, etc.1 availability of FI accessories.Although on-line options are On-line pre-reduction is therefore a challenging and diYcult not as numerous as oV-line separation/enrichment procedures,1 task, as demonstrated in several recent publications on there are several viable coupling schemes which are depicted As,52,53,71,95,116–122 Se69,96,99–102,123–126 and Te.127 Their main in Fig. 3. They are listed below in an approximate order of advances and problems are briefly summarised below. their decreasing importance. For arsenic, the best results are obtained with L-cysteine (a) The VG-ETAAS technique is based on in-atomiser (2–4% m/v) in dilute mineral acid (10–50 mmol l-1 HCl or trapping of hydrides or other vapours for subsequent quantifi- HNO3) at 70–100 °C.52,53,95,118,120,122 Arsenate can be reduced cation.It is developing very rapidly (Fig. 1), gaining advan- within 1 min, depending on the temperature and concentration tages from both VGAAS and ETAAS, and undoubtedly is the of L-cysteine, showing good tolerance to changes in pH128 and most successful hyphenation approach in VGAAS.HCl concentration.53 Monomethylarsonate (MMA) and (b) CVAAS with amalgamation is already a classical and dimethylarsinate (DMA) are also pre-reduced to their tervalent state, forming organosulfur derivatives (thiolate com- very successful technique for determination of Hg down J. Anal. At. Spectrom., 1999, 14, 147–162 151ETAAS or FI-MWD-HG-ETAAS for ‘first-order speciation’ of As in urine.93 VG-ETAAS can be adapted to most ETAAS instruments with minor hardware modifications.A quartz capillary positioned on the GA autosampler arm is the best interface for introducing hydrides into the pre-heated GA. (c) The absolute sensitivity of VG-ETAAS approaches the ETAAS figures for characteristic masses, with peak-height (Ap) and integrated absorbance (Aint) measurements (mp and m0, respectively) being within a few pg and <100 pg, while the relative sensitivities and real LODs are also dependent on sample volume and reagent blanks—LODs down to a few ng l-1.Generally, the sensitivity gain over direct HGAAS or ETAAS is within a factor of 10–100 and the relative LODs are improved by at least one order of magnitude (see Table 5). Enrichment from larger sample volumes (>1 ml ) is more Fig. 3 Schematic representation of possibilities for on-line enrichment and separation in VGAAS, with anwn on the right. conveniently performed by CF or the batch mode of VG. (d) Better interference control compared with direct HGAAS is provided:153,204–206 the liquid-phase chemical interferences are to ng l-1 levels—well documented and straightforward in oper- decreased owing to higher dilution factors of samples/digests and ation, fully automated and widely commercialised; for more the FI mode of operation with Aint measurements,153 while the data, see recent monographs.9,35,36 gas-phase interferences which are common in HGAAS with QTA (c) VG-CT-AAS relies on enrichment by cryogenic trapping are eliminated to a large extent owing to the electrothermal and subsequent volatilization for AAS determination with FIT atomisation mode.153,160,175,190,204,205 Still, huge amounts of other or QTA.It has found many early applications1,32,33 for on-line HFEs may be depressive,153,175 especially with large sample preconcentration and speciation, mostly in the batch VG mode volumes,175 low modifier (Pd) mass153 and Ap mode of from large sample volumes (10–100 ml ), thus reaching LODs measurement.153 of around a few ng l-1 for As, Ge, Hg, Sb, Se and Sn.Whereas (e) Hydrides and Hg0 vapours are best trapped on graphite this approach is more diYcult to automate and bulky and surfaces which are chemically modified with noble metals such more expensive to run for enrichment only, it currently rep- as Pd (0.02–0.1 mmol )27,202,204,205 or Ir (0.02–0.25 resents interest mainly for speciation by VG-CT/GC-AAS mmol ).66,175,205 The scarce comparative studies give preference (FIT) of hydride-forming, ethylated or otherwise derivatized to Pd over other noble metals (e.g., Pd>Pt>Rh>Ru27 or inorganic and organoelement species (As, Hg, Sb, Sn)—see Pd>Ir, Pt204) owing to some positive assets of this common Section 6.2 and the recent discussion on HG-CT-AAS and chemical modifier as summarised elsewhere:200 (i) high trap- HG-CT/GC-AAS by De¡dina and Tsalev1 and Howard.130 ping eYciency; (ii) relatively low collection (trapping) tempera- (d) On-line FI-ion-exchange–VGAAS has found sporadic tures; (iii) small modifier mass (<10 mg); (iv) improved applications for preconcentration67–71,142–152 or on-line ETAAS performance characteristics for volatile elements, removal of interfering matrix components72–74 in the determi- resulting in lower mp; and (v) more uniform and less critical parameters of thermal programming (collection, pyrolysis and nations of As,67,70,71,142,143 Bi,144 Hg,112,114,145–147 Sb,148,149 atomisation temperatures, Tcoll, Tpyr and Tat, respectively) for Se68,69,123,144,149,150 and Sn.151,152 A promising approach diVerent analytes, which greatly facilitates the multi-element among these techniques is the immobilization of tetrahydrobosequestration and quantification.rate (BH4-) on an anion exchange resin in a microcolumn for Physical modification of the graphite tube (GT) by mechan- enrichment and quantification of As,67,142 Sb149 and Se68,149 ical removal of the pyrolytic graphite layer so as to expose a in an automated FI-HGAAS system.rough, more reactive graphite surface may be acceptable in some cases, e.g., with unstable hydrides of Bi,24 Cd,157 5.1 VG-ETAAS Sb21,24,174 and Te188 or for other practical considerations. VG-ETAAS has been applied to various real samples at Thus, applying modifier injections before every HG-ETAAS analyte levels in the mg l-1 and sub-mg l-1 concentration range. run has exhibited some drawbacks, as noted by Shuttler Table 5 gives a guide to the vast literature, which has been et al.:199 complications in the existing hardware and software confined here to selected recent studies utilising modern and increase in the total cycle time when pipetting (Pd) ETAAS technology and chemically modified graphite atomiser modifier solution, then thoroughly washing the autosampler (GA) surfaces. Some important assets, results and problems capillary, performing drying/pyrolysis for conditioning of the of this important technique are discussed below (see also a modifier (200–1200 °C), then cooling the GA to the optimum recent monograph1 and review19).Tcoll, and finally introducing hydrides through the same capil- (a) The scope of VG-ETAAS has been steadily expanding lary; or alternatively, applying tedious manual pre-injection of in the last decade to comprise all HFEs (except for Tl, which modifier solution before each run. Therefore, ‘permanent is very problematic in HG1); cold vapour of Hg0;28–30,165–167 modification’199 of the GT or platform surface with a less ethylated derivatives40 of Pb,25,170 Se49,178,179,184 and Sn; alkyl volatile—ideally, non-volatile—modifier such as noble metal hydrides of As,93,154,200,201 Pb171 and Sn;23,183–185,200,201 CrIII (Pd–Ir;199 Ir66,175), carbide (Zr-168 or W-treated surfaces186) b-diketonates,203 etc.; thus oVering some (still modest) speci- or a noble metal on carbide coating (Ir–Zr or Ir–W)131,200,201 ation possibilities at sub-mg l-1 levels for As,93,131,154,201 Ge,164 has proved an eYcient and practically useful approach.Pb,171 Se,178,179,182,201 Sn183–185,201 and Te.188 HG-ETAAS has Comparative studies of noble metals versus carshown very competitive performance characteristics for bides158,162,166,169,198,201 and of platform versus wall trapping/ Ge,158–164,191,202 since its hydride (GeH4) cannot be otherwise atomisation161,162,170,175,184,186,199 are not necessarily in agreeatomised in QTA or FIT, while flame atomisation lacks ment because of diVerences in the apparatus and chemistry sensitivity for Ge.1 involved (see Table 5 for individual analytes and specific (b) FI-VG-ETAAS is now fully automated and conditions).Iridium may be preferred for better sensicommercialised, 66 oVering sampling frequencies of 20–90 h-1 tivity,156,166,169,198 lower Tcoll 198,201 or both,201 while Zr-treated and possibilities for other promising hyphenations, e.g., with surfaces (1–3 mmol of Zr) provide longer lifetimes, e.g., 400 firings.162,186,197,198 The merits of the stabilised temperature on-line ion exchange enrichment (for Sn152) or FI-UV-HG- 152 J.Anal. At. Spectrom., 1999, 14, 147–162Table 5 Selected application studies on vapour generation-ETAAS with in-atomiser trapping (VG-ETAAS) Analyte Sample Trapping surface/modifier Brief comment Ref. As Water (river, sea) Pd FI-HG-ETAAS; Tcoll 200–600 °C; 153 LOD 36 pg As Aquatic plant, biological Ir-treated GT (100 mg Ir) On-line UV photooxidation or MWD 93 tissues, urine, water with FI-HG-ETAAS for ‘first-order’ speciation; LOD 0.14 ng As Water ( lake, river) 2 mg Pd dried at 100 °C on GT Speciation of seven As species after 154 wall oV-line ion chromatographic separation; CF-HG-GFAAS; Tcoll 600 °C; Tat 2400 °C; LODs 1.6–1.9 mg l-1 for i-AsIII, i-AsV, MMA, DMA, AB, AC and paminophenylarsenic acid Cd Aqueous solutions 5 mg Pd VG-ETAAS; Tcoll 150 °C; Tpyr 800 °C; 155 LOD 60 ng l-1; STPF Cd Sea-water Ir-, W- and Zr-treated GTs FI-VG-ETAAS; Tcoll 20 °C; Ir coating 156 recommended; m0 3 pg; LOD 4 ng l-1 Cd Environmental CRMs: lobster, Uncoated GT Batch or CF-VG-ETAAS in the 157 plant, sediment, sea-water, presence of thiourea, CoII and soil didodecyldimethylammonium bromide; Tcoll 200 °C; Tat 1800 °C; LOD 10 pg Ge Aqueous solutions Pd-, Zr- and Pd–Zr-coated GTs Sample injection versus HG-ETAAS; 158 Tcoll 800 °C Ge Coal fly ash, river sediment, Pd-treated GT HG-ETAAS; 0.15–0.6 mol l-1 HClO4 159 soil medium for HG; m0 46 pg; LOD 36 pg Ge Garlic, geological CRMs, Pd-treated GT FI-HG-ETAAS; Tcoll 400 °C; LOD 160 ginseng, tap water 4 ng l-1 Ge NaCl Pd, Pd–Mg(NO3)2 Direct ETAAS and HG-ETAAS 161 compared; Tcoll 800 °C; Tpyr 1200 °C; LOD 30 pg Ge Rock, sediment, steel Pd, Ir, Ir–Pd, Pd–Mg(NO3)2; HG-ETAAS; Tcoll 400–500 °C (noble 162 permanent coatings: Nb, Ta, W, metal ) or 500–600 °C (carbide); mp Zr 10–54 pg; Zr-treated GT or platform recommended Ge Aqueous solutions Pd-treated versus electrolytically FI-HG-ETAAS; Tcoll 400 °C; Tat 163 coated GTs 2400 °C Ge b-Carboxyethylgermanium Pd-treated GT OV-line cation exchange 164 sesquioxide chromatographic separation; digestion with K2S2O8; HG-ETAAS; Tcoll 800 °C; Tat 2600 °C Hg Water Ir permanent modifier or Au–Pt FI-CV-ETAAS; Tcoll 150 °C; LOD 5.2 28 gauze or 4 ng l-1 in 25 ml, respectively Hg Water 120 mg Ir permanent modifier CV-ETAAS; Tcoll 150 °C; Tat 900 °C; 30 lifetime of coating 500 firings; LOD 70 pg; Zeeman STPF Hg Water Pd-treated GT CV-ETAAS; Tcoll 250 °C; m0 114 pg; 165 LOD 31 pg Hg Water GT with Au–Pt gauze insert Automated CF-CV-ETAAS; Tat 29 750 °C; LOD 15 pg or 1.5 ng l-1 in 10 ml Hg Water (fresh, sea) Ir-, W- and Zr-treated GTs CV-ETAAS; Ir preferred; m0 240 pg; 166 LOD 60 ng l-1 In Aqueous solutions Pd-treated platform HG-ETAAS; Tcoll 800 °C; mp 0.63 ng 167 Pb Coal fly ash, mussel, tea, water Zr-treated GT HG-ETAAS; Tcoll 300 °C; m0 52.8 pg, 168 LOD 242 pg Pb Sediment Ir, Pd–Ir, Ir–Mg, W and Zr HG-ETAAS; radiotracer 209Pb study; 169 coatings best sensitivity with Ir; m0 21 pg; LOD 0.25 ng Pb Water (riverine, sea), fish, 150 mg Ir (conditioned at 1100 °C) FI ethylation with NaB(C2H5)4; 170 lobster tetraethyllead trapped at 300 °C; LOD 12 pg or 2 ng l-1; Zeeman STPF Pb Aqueous solutions GT with integrated platform Generation and trapping of Me2, Me3, 171 (THGA) Et2 and Et3 derivatives studied; HG-ETAAS; Tcoll 350 °C; Tpyr 600 °C; Tat 1600 °C; incomplete isoformation Sb Sediments 2 mg Ir on Zr-treated platform FI-HG-ETAAS after bomb 172 decomposition Sb Wine Pd-treated GT Batch HG-ETAAS in the presence of 173 thiourea; Tcoll 300 °C; Tat 2400 °C; LOD 39 pg or 0.13 mg l-1 Sb Water Uncoated GT FI-HG-ETAAS; Tcoll 500 °C; LOD 174 15 pg or 5 ng l-1 J.Anal. At. Spectrom., 1999, 14, 147–162 153Table 5 (Continued) Analyte Sample Trapping surface/modifier Brief comment Ref. Se Aqueous solutions 50 mg Ir permanent modifier FI-HG-ETAAS; STPF 175 Se Nutritional supplement 50 mg Ir (permanent modifier) FI-ETAAS; Tcoll 250 °C 176 formula Se Oral nutrition liquids Pd-treated GT Selenosugar decomposed by K2S2O8 at 177 85 °C for 15 min; i-SeIV determined in a separate aliquot; CF-HGETAAS; Tcoll 700 °C; Tat 2400 °C; m0 21 pg; LOD 36 pg Se Soil Pd HG-ETAAS; Tcoll 700 °C; (CH3)2Se 178 and (C2H5)2Se also determined; m0 15 pg Se Soil Pd-treated GT GC-trapping-ETAAS for speciation of 179 (CH3)2Se, (C2H5)2Se and (CH3)2Se2 with mp 18, 12 and 30 pg, respectively; Tcoll 500, 700 and 700 °C, respectively Se Urine PdCl2 (dried at 120 °C) HG-ETAAS; Tcoll 400 °C; LOD 180 20 ng l-1 or 20 pg Se Urine 120 mg Ir FI-HG-ETAAS after oV-line digestion 181 with Br-–BrO3-; Tcoll 250 °C; LOD 3 mg l-1 Se Industrial sewage sludge Ir-treated platform (150 mg) FI ethylation with NaBEt4-trapping 49 of Et2Se-ETAAS; selective determination of SeIV; Tcoll 750 °C; Tat 2000 °C; LOD 0.05–0.08 ng Se Sea-water Ir-treated platform Speciation protocol: SeIV determined 182 without pre-treatment; total Se after pre-reduction in 5 mol l-1 HCl at 100 °C or UV photolysis at pH 10 for 30 min; FI-HG-ETAAS; Tcoll 300 °C; LOD 1.5 ng l-1 Sn Aqueous solutions 100 mg Ir permanent modifier FI-HG-ETAAS for i-Sn and butyltins; 183, 184 Tcoll 500 °C; complete isoformation of species not possible; STPF Sn Geological CRMs, hair, Pd-treated GT FI–on-line enrichment on strongly 152 serum, basic anion exchanger with tap water HG–ETAAS; LOD 10 ng l-1; 30 h-1 Sn Sea-water Zr-treated GT HG-ETAAS; Tcoll 500 °C; m0 14 and 185 20 pg for i-SnII and Bu3Sn+, respectively; Bu3Sn+ speciation after extraction in CH2Cl2 Sn Low-alloy steel Nb-, Ta-, W- and Zr-treated GTs HG-ETAAS; Tcoll 600 °C; LOD 186 or platforms 25 pg; m0 17 and 20 pg for Zr- and W-treated GT, respectively (preferred for better sensitivity and long-term stability) Sn Liver, plant, sediment, steel Pd-treated platform HG-ETAAS; Tcoll 300 °C; Tat 2300 °C; 187 LOD 66 pg or 7 ng l-1 Te Aerosols, garlic, sediment, Uncoated GT Batch HG-ETAAS; speciation of i-TeIV 188 water and total Te; Tcoll 110 °C; Tat 2000 °C; LOD 2–4 pg As, Sb Rye grass CRM Pd-treated GT (10 mg) CF-HG-ETAAS; Tcoll 120–700 and 189 120–500 °C; m0 69 and 57 pg; LOD 25 and 36 pg, respectively As, Sb Aqueous solutions Pd-treated GT (10 mg Pd treated CF-HG-ETAAS compared with 190 at 1200 °C) ETAAS and CF-HGAAS and HG-ICP-AES; multi-element capabilities discussed As, Ge HCl (for As), sea-water (As, 10 mg Pd HG-ETAAS; Tcoll 200–800 and 400– 191 Ge), urine (As) 900 °C As, Se Aqueous solutions Zr-treated versus Pd-treated GTs HG-ETAAS; Tcoll 800–900 and 600– 192 800 °C; mp 43 and 77 pg, respectively As, Se Mineral waters 4 mg Pd HG-ETAAS; Tcoll 300 °C; Tpyr 450 °C 193 As, Se Sea-water Pd–Ir-treated GT Electrochemical HG-ETAAS; Tcoll 194 350 °C; Tat 2400 °C; LOD 84 and 75 pg Bi, Se Aqueous solutions Ir-treated platform (100 mg) Simultaneous FI-HG-ETAAS; Tcoll 195 250 °C; Tat 2200 °C; m0 55 and 41 pg; LOD 0.13 and 0.19 mg l-1, respectively Se, Te Garlic, mussel, rice flour Ag-treated versus Pd-treated GTs HG-ETAAS; Ag preferred for lower 196 Tat (1800 °C); Tcoll 200–400 and 200–800 °C; m0 17 and 18 pg, respectively 154 J.Anal. At. Spectrom., 1999, 14, 147–162Table 5 (Continued) Analyte Sample Trapping surface/modifier Brief comment Ref.Se, Te Low-alloy steel Ir, Ir–Mg, Ir–Pd; carbide-treated FI-HG-ETAAS and CF-HG-ETAAS; 197 surfaces (Nb, Ta, W, Zr) Tcoll 600 °C; m0 11 and 12 pg; LOD 11 and 7 pg on Ir coating in FI and CF modes, respectively As, Bi, Sb Low-alloy steel Ir, Ir–Mg, Ir–Pd; carbide-treated FI-HG-ETAAS with in situ trapping 198 surfaces (Nb, Ta, W, Zr) on permanently modified GT or tube with integrated platform; Zr preferred for longer lifetime over 400 cycles; m0 16, 9 and 15 pg; LOD 15, 27 and 10 pg, respectively As, Bi, Se Aqueous solutions 50 mg Pd–50 mg Ir FI-HG-ETAAS; permanent modifier 199 useful up to 300 cycles; m0 45, 76 and 61 pg As, Bi, Sb, Sn Aqueous solutions 2 mg Ir on Zr-treated platform HG-ETAAS with permanent 131, 200 modification; eVect of L-cysteine on inorganic and organoelement species studied; multi-element prospects addressed As, Bi, Sb, Se, Sn, Te Sea-water, sediment 2 mg Ir on Zr- or W-treated Permanent modification; HG-ETAAS; 200, 201 platform trapping and stabilisation of organo-As, Sn and Se species studied; long-term stability during 600–700 thermal cycles; Ir–Zr preferred Fig. 4 Schematic representation of possibilities for on-line coupling of chromatography with VGAAS for speciation analysis, with analytes shown on the right. platform furnace (STPF) technology207 are well recognised in Te197 and Ge,162 respectively; (iii) appearance of double peaks under some unfavourable conditions,158,162,171,175,197,198,200,210 ETAAS; accordingly, the trapping on and atomisation oV carbide-treated platforms have proved advantageous with e.g., because of partial deposition of the analyte on the tube wall rather than on the platform or due to inhomogeneous some diYcult analytes such as Ge161,162 and Sn.183,184,186 Mechanisms of hydride trapping and analyte stabilisation have distribution of the (mixed) modifier on the trapping surface, thus yielding diVerent atomisation sites;158,209,210 (iv) species- been discussed in more detail elsewhere,1,24,27,158,167,208–210 and chemical modification in ETAAS and VG-ETAAS has been dependent Aint responses, i.e., incomplete isoformation, of some organoelement hydrides (Sn183–185,201) or alkylated derivatives reviewed recently by Tsalev et al.211,212 (f ) The multi-element capabilities of HG-ETAAS compared (Pb;171 Se179) because of diVerent optimum conditions for their VG, collection and atomisation.with direct HGAAS are now better recognised, owing to the availability of multi-element ETAAS instrumentation195 and VG-HGAAS has been reviewed earlier by De¡dina and Tsalev1 and Matusiewicz and Sturgeon.19 more easily attained compromise conditions of HG and ETAAS, as demonstrated with some analyte pairs such as As–Sb,190 As–Se192 and Bi–Se195 and as shown elsewhere by Tsalev et al. in the absence (As–Bi–Te; As–Bi–Se)201 and 6 Hyphenated techniques for speciation analysis presence (As–Bi–Sn; As–Bi–Sb)131 of L-cysteine.(VG-CT/GC-AAS, HPLC-VGAAS, HPLC-UV- (g) Some problems with the VG-ETAAS technique have been VGAAS, HPLC-MWD-VGAAS, etc.) encountered: (i) pronounced reagent blanks for As, Cd, Pb and Sn; (ii) gradual accumulation of pyrolysis products on the tip Speciation analysis is a modern, important application area of VGAAS with QTA or FIT atomisation.1–6 Other AAS options of the quartz capillary,201 which may entail an ‘adsorptive carry-over eVect’ at high collection temperatures, e.g., above such as flame AAS (FAAS) or ETAAS are virtually out of consideration as chromatographic detectors because of their 350, 450, 450, 600, 600 and 650 °C for Bi,198 Pb,169 Sb,198 Sn,186 J.Anal. At. Spectrom., 1999, 14, 147–162 155Table 6 Typical examples of VG-GC-AAS procedures Analyte Sample Brief outline of procedure and species determined Ref. As Water (river) CF-HG-CT/GC-AAS (QTA): i-As (arsenite+arsenate), MMA and DMA derivatized with L- 47, 128 cysteine to their respective complexes; HG; cryotrapping of AsH3, MeAsH2 and Me2AsH on OV-3 coated Chromosorb W-HP (15%) at -196 °C; fractional vaporization of arsines; QTA at 900 °C; LOD 30–98 pg or ~50 ng l-1 Hg Aquatic plant Batch VG-CT/GC-AAS (QTA): samples extracted with 0.1 mol l-1 HCl under sonification; 78 i-HgII, MeHg+, Me2Hg and Et2Hg derivatized with NaBH4; volatile products cryotrapped, then fractionally volatilized; QTA at 825 °C; LOD 50–110 pg (as Hg) Hg Fish Benzene extraction in the presence of CuII–Br-–1.5–2 mol l-1H2SO4; CGCa–AAS with 75 on-line pyrolysis at 700 °C and QTA at room temperature; Me-, Et-, Et2- and Ph-mercury species determined with LOD of ~0.1 ng or 0.3 mg g-1 Hg Fish, marine tissue Batch VG-CT/GC-AAS (QTA): solubilization with methanolic KOH; derivatization with 47 NaBEt4; QTA at 830 °C; LOD 4 and 75 pg Hg for MeHg+ and labile Hg2+, respectively Hg Water Enrichment (2×103×) by batch sorption on dithiocarbamate resin; elution with acidic 32 thiourea (pH 1.0); dithiocarbamate extraction into hexane; butylation with Grignard reagent; GC-AAS (QTA) with on-line pyrolysis at 800 °C; LOD 0.03 and 0.4 ng l-1 for MeHg+ and i-Hg, respectively Pb (Rain) water CGC-AAS (FIT) after in situ butylation with tetrabutylammonium tetrabutylborate; 220 extraction into pentane; drying; hexane added; evaporation; 103×enrichment; 0.1–10 ml injections; H2 and air added to QTA at 900–1000 °C; LOD 13–30 pg for Me4-, Me3Et-, Me2Et2-, Me3Et-, MeEt3-, Me3Pe-, Et4-, Me2Bu2-, Et3Bu-, Et2Bu2- and Bu4-lead species Sn Algae,221 marine tissue,221 Solid samples extracted with acetic acid; batch HG-CT/GC-AAS with FIT (H2 and O2 221, 222 sediment,221,222 water221 introduced); interference eVects addressed; Bu-, Bu2- and Bu3-tin species determined Sn Fish Extraction with methanolic 0.5 mol l-1 HCl under microwave heating or sonification; 50 ethylation with 1% m/v NaBEt4; CT/GC-AAS with FIT (H2 and air added); LOD 2–4 ng g-1 (as Sn) for Me-, Me2-, Me3-, Bu-, Bu2- and Bu3-tins Sn Fish,223 sediment,224 Batch HG-CT/GC-AAS (FIT); tissues enzymatically hydrolysed; sediments extracted with 223–225 shellfish,223 water225 acetic acid; hydrides trapped on OV-101 coated Chromosorb GA W-HP (10%); FIT atomisation (H2 and air added, as detailed in ref. 225); LOD 1.8 ng g-1 (as Sn) for Bu3Sn+ in sediment224 and ng g-1 levels for Bu-, Bu2- and Bu3-tins in tissues223 Sn Fish, sediment Microwave assisted leaching of butyltins from sediments with 50% CH3COOH or 51 solubilization of tissue with Me4NOH; derivatization with NaBEt4 with extraction of products into hexane; CGC-AAS with FIT (H2 and air added); LOD down to 5 ng g-1; 5 h-1 Sn Sea-water Batch HG-CT/GC-AAS (FIT): HG at pH 2.0; cryotrapping on OV-3 coated Chromosorb W 33 AW-DMCS for i-Sn, Me-, Me2-, Me3- and Bu-tin hydrides or on silanized glass-wool for i-Sn, Bu-, Bu2- and Bu3-tin hydrides; LOD ~50 pg in 100 ml Sn Urine Bu-, Bu2- and Bu3-tin species chelated with tropolone and extracted in diethyl ether; batch 226 HG; hydrides extracted into hexane; CGC-AAS (FIT) (H2 and air added); LOD 1–2 mg l-1 Sn Water Ionic methyltin and butyltin extracted into pentane as dithiocarbamate derivatives at pH 5; 227 evaporation; derivatization with pentyl-Grignard reagent (PeMgBr); extraction in octane; GC-AAS; FIT atomisation (H2 and air added); LOD 0.16–0.40 ng Sn or 4–10 ng l-1 aCGC=capillary gas chromatography. low sensitivity (FAAS) or discontinuous mode of operation and FI-HG-ETAAS93,131,201 and/or on-line treatments are (ETAAS; VG-ETAAS).Attractive features of VGAAS are its involved: digestion,71,93,100 pre-reduction69,71,119,120,123–125 and element-specific response, widespread availability, well estab- separation.68,69,71,123 This approach is attractive for its simlished methodology, sensitivity down to sub-nanogram levels plicity, high sample throughput (30–60 h-1) and relatively low and good performance for several elements with biological costs, but its scope is limited and a lower grade of speciation and environmental importance: As, Hg, Pb, Se, Sn, etc.35,36 information is obtained—a ‘first-order speciation’.93 Some Some limitations of VGAAS coupled techniques are listed in typical tasks are given below with brief comments (see also Table 2; VGAAS is now seriously rivalled as a chromato- selected application examples in Tables 4 and 5).graphic detector by other sensitive and commercially available (a) Redox speciation, i.e. diVerentiation between the techniques: ICP-MS,213–215 atomic fluorescence spectrometry common oxidation states of the inorganic (i) analyte species (AFS)216–218 and microwave induced plasma (MIP)- such as i-SeIV (selenite)177,182,201 and i-TeIV (tellurite)188 by AES,216,217,219 which are able to provide better LODs (>10×), FI-HG-ETAAS177,182,188,201 or FI-ion-exchange-HGAAS broader working range and less chemical treatments, but with (QTA)68,69 down to LODs of a few ng l-1.The total amount more expensive hardware and running costs. is then determined in another sample aliquot after a suitable Some possibilities of coupling VGAAS with chromato- treatment, and the contents of the i-SeIV and i-TeIV fractions graphic methods are presented schematically in Fig. 4 and are evaluated ( less precisely) by diVerence. A similar approach selected application examples of hyphenated VG-GC-AAS can be applied for i-AsIII (arsenite) and i-SbIII (antimonite) and HPLC-HGAAS procedures are given in Tables 6 and 7, utilising the eVect of pH/reaction medium on respectively.Further discussion will be confined to three HG,1,119,120,237,238 but the discrimination is not as reliable essential speciation approaches: (i) automated on-line (FI ) because of interferences from methylated species and other non-chromatographic speciation; (ii) hyphenated VG-GC- problems.1 AAS; and (iii) HPLC-HGAAS coupling.(b) Some kind of speciation information can be obtained by alternating the ‘on–oV’ operation modes for on-line pre- 6.1 On-line speciation by non-chromatographic HGAAS reduction69,71,100,119,123–125,182 or decomposition,71,100,130,182 techniques assisted by microwave heating71,93,100,125 or UV irradiation.18,93,130,182 Thus a ‘first-order’ diVerentiation In this group of speciation analyses are utilised the speciesspecific diVerences in HG patterns in FI-HGAAS1,119,120,132,237 between the total amount and the ‘hydride-forming’ fraction 156 J.Anal. At. Spectrom., 1999, 14, 147–162Table 7 Typical examples of HPLC-HGAAS procedures Analyte Sample Brief outline of procedure and species determineda Ref. As Aqueous solutions,18 HPLC-UV-HGAAS (QTA); strong anion exchanger SAX column; 166 ml injections; 18, 142 sediment extract142 MP 30 mmol l-1 KH2PO4, pH 4.6; UV photooxidation with 2% m/v K2S2O8–1% m/v NaOH; QTA at 900 °C (air added); LODs 50 pg or 0.3 mg l-1 (as As); arsenite, arsenate, MMA, DMA, AB, AC, o- and p-arsenilate, phenylarsonate and Me4As+ studied As Coal fly ash, sediment, HPLC (IC)-HGAAS (QTA); Dionex PAX-500 anion exchange column; 250 ml 228 soil injections; MP 30 mmol l-1 NaOH–1% v/v MeOH; LOD 200 pg or 0.8 mg l-1 in extracts for arsenite and arsenate As Fish,94 urine,91 water91,94 HPLC-MWD-HGAAS (QTA)91 or HPLC-HGAAS with on-line thermo-oxidation;94 91, 94 PRP X-100 strong anion exchanger column; 100 ml injections; MP 17 mol l-1 NaH2PO4–Na2HPO4 buVer, pH 6.0; oxidant 5% m/v K2S2O8–5% m/v NaOH; LOD 0.5, 0.8, 0.6, 0.6, 0.4 and 0.4 ng As for arsenite, arsenate, MMA, DMA, AB and AC, respectively As Marine tissue CRMs HPLC-MWD-HGAAS (QTA): enzymatic extraction with 5 g l-1 trypsin–0.1 mol l-1 95 NH4HCO3; BAX-10 strong anion exchanger column; on-line MWD with 5% m/v K2S2O8–3.4% m/v NaOH in a 15 m KR at 50 W; on-line pre-reduction with 2.5% L-cysteine–30 mmol l-1 HNO3 at 85 °C; antifoam added; LOD 0.5, 0.6, 2.5, 3.3, 5.3 and 5.9 mg l-1 for ‘reducible As’, ‘total As’, AB, MMA, DMA and i-As, respectively As Sea-food MMA and DMA by HPLC-HGAAS (QTA)229 and AB by HPLC-MWD-HGAAS 92, 229 (QTA)92 determined with LODs down to 0.2, 0.3 and 0.68 ng g-1 for MMA, DMA and AB, respectively; strong anion exchanger column PRP X-100; MP phosphate buVers; MWD in the presence of 1% m/v K2S2O8–2.5% m/v NaOH As Serum 1+1 dilution with water; ultrafiltration; HPLC-HGAAS (QTA); 300 ml injections; 230 strong anion exchanger or RP column; MP phosphate buVers; LOD 0.49, 0.44, 0.92 and 0.40 mg l-1 for arsenite, arsenate, MMA and DMA, respectively As Serum Serum (0.9 ml) deproteinized with 0.1 ml of 4 mol l-1 HClO4; HPLC-UV-HGAAS 90 (QTA); cation exchange column (Dionex Ionpac CS 10); MP 50 mmol l-1 NaH2PO4–100 mmol l-1 HCl, pH 1.6; post-column Ar segmentation of flow; UV photooxidation in the presence of 1.5% m/v K2S2O8–4% m/v NaOH; LOD 1.0, 1.3, 1.5 and 1.4 mg l-1 (as As) for MMA, DMA, AB and AC, respectively As Urine HPLC-UV-HGAAS (QTA); strong anion exchanger column PRP X-100 or SAX 1; 64 50 ml injections; gradient elution with phosphate buVers (KH2PO4–K2HPO4); UV photooxidation (93 s) with 4% m/v K2S2O8–1 mol l-1 NaOH; LOD 2–6 mg l-1; arsenite, arsenate, MMA, DMA, AB, AC, TMAO and Me4As+ studied As Urine,231,232 HPLC-HGAAS (QTA) with various strong anion exchanger70,121,231 or RP column 70, 121, 231–234 water70,121,232–234 (ion pair HPLC);232–234 LOD down to 0.1–0.3 ng or ~1 mg l-1 (as As) for arsenite, arsenate, MMA and DMA Sb Water HPLC-HGAAS (QTA); PRP X-100 strong anion exchanger column; 100 ml 235 injections; MP 2 mmol l-1 phthalic acid, pH 5; LOD 5 and 0.6 ng for i-SbIII and i-SbV, respectively Sb Water HPLC-HGAAS (QTA); Supelcosil LC-SAX 1 strong anion exchanger column; 236 200 ml injections; MP 50 mmol l-1 ammonium tartrate buVer, pH 5.5; LOD 2 and 1 mg l-1 for i-SbIII and i-SbV, respectively Se Water, urine HPLC-MWD-HGAAS (QTA); vesicle mediated HPLC with a RP C18 column; 100– 101, 102 500 ml injections; gradient elution; on-line MWD with BrO3-–Br- at 45 W (1 min); LOD down to 1 mg l-1; selenite, selenate, selenomethionine, selenoethionine and selenocystine studied Se Water (tap) HPLC-MWD-HGAAS (QTA); PRP X-100 strong anion exchanger column; 100 ml 96 injections; MP 100 mmol l-1 NaH2PO4–Na2HPO4 buVer, pH 6.8; on-line MWD with 3% m/v K2S2O8–2% m/v NaOH at 650 W; cooling; on-line pre-reduction with 10 mol l-1 HCl on microwave heating; cooling; LOD 1.4, 2.2 and 1.1 ng for selenite, selenate and Me3Se+, respectively aAbbreviations: MP, mobile phase; IC, ion chromatography; RP, reversed phase; KR, knitted reactor.can be obtained for As71,93,130 and Se.69,100,124 Admittedly, Sn,1,5,36 and are close to routine applicability for Hg, Pb and Sn.these fractions are not very precisely defined and may consist Most of these procedures require manual, labour-intensive of several analyte species. For example, the ‘toxicologically sample preparation steps such as solubilization, derivatization of relevant arsenic’ in urine is easily determined by the inorganic and organoelement species to volatile products, FI-HGAAS129,132 or FI-HG-ETAAS,93 being practically reppre- and post-derivatization enrichment (103×), clean-up and resented by the ‘hydride-forming’ fraction, i.e., the sum of drying of organic extracts, etc.These steps are prone to volatiliz- arsenite+arsenate+MMA+DMA;35,36 hence the sum of ation, retention and transformation losses and also incomplete, these four species can be determined and discriminated from variable and species/matrix-dependent recoveries. Hence prefer- the non-toxic arsenic derived from seafood (AB, Me4As+, ence is given in this discussion to those VG-GC-AAS approaches arsenosugars, etc.) with sampling frequencies as high as (see Fig. 4 and Table 6) which rely on in situ derivatization in 50–55 h-1.132 aqueous media and on-line enrichment (approximately ranked in performance): (i) ethylation40 for Hg46,47 and Sn;50,51,239–242 6.2 Coupling with GC (ii) hydridization1 for As,32,47,128–130,243–246 Hg78 and GC-AAS techniques31 have been extensively documented Sn;33,221–226,247 (iii) cryogenic trapping (CT) or CT/GC on (>100 papers) and reviewed.1–6,130,216 They are best estab- silanized chromatographic supports with subsequent fractional volatilization, viz., VG-CT-AAS (FIT) or VG-CT/GC-AAS lished in speciation determinations of As, Hg, Pb, Sb, Se and J.Anal. At. Spectrom., 1999, 14, 147–162 157(FIT)130 (see Table 6); and (iv) in situ derivatization-HG-head- post-column merging flows and multi-step manipulations; (v) reagent blanks. Some possible remedies for improving the space (HS)-GC-AAS (FIT).245,246 relative sensitivities, at the expense of other compromises and Recent research has brought some progress in the (solid ) side-eVects, have been tried: forcing the injection volume up sample preparation step, owing to sonification239 and microto 200–350 ml with favourable matrices101,102,228,236 but com- wave-accelerated leaching or solubilization,51,242 e.g., resulting promising the chromatographic resolution and lifespan of the in an impressive, one-step integrated solubilization, extraction pre-column/column; post-column segmentation of flow;34,90 and ethylation of organotins from marine tissues in a focused, optimisation of the gas–liquid separator (GLS) construction low-power microwave field (20–80W) within 3 min.242 and size;58,60,64,65,101,233 employing shorter columns;236 increas- 6.3 Coupling with HPLC ing the NaBH4 concentration in the HG step for AsV with post-column UV photooxidation;64,65 resorting to a pre-HPLC Generally, HPLC-VGAAS is the most versatile hyphenated enrichment step,70,110–114,232 etc.AAS technique for speciation. Attractive features of the HPLC-VGAAS coupling are: (i ) good separation potential, 6.3.2 Hardware complexity. The grade of hardware even with non-volatile and unstable species; (ii) minimised sophistication and complexity of chemical treatments—generpre- column sample handling; (iii) working with aqueous ally increasing in the order given in Fig. 4—result ultimately sample solutions/leachates but also with mixed solvents [modi- in some systems that are operated properly only by expert’s fied mobile phases in reversed-phase (RP) chromatography], hands and eventually are becoming somewhat bulky and if required; (iv) large selection of HPLC columns and chemistr- diYcult to control reliably. As many as 15–20 instrumental ies (anion- or cation-exchange, RP, ion-pair, etc.); (v) small and chemical parameters call for careful optimisation and sample size requirements (typically 50–100 ml injections); (vi) proper control during long-term operation.Safety precautions well established HPLC instrumentation and methodology; and in working with the TS–THG interface79–89 and ( less so) with (vii ) automation. Research eVorts in this field show remarkable the FIT atomiser and also handling of copious amounts of dynamics (see Fig. 1). Speciation procedures are best elabor- toxic waste eZuents cannot be overlooked.ated for As: without70,122,228–234,248–250 or with post-column Interfacing HPLC with VGAAS (QTA) is technologically digestion, thermally94,121 or microwave-assisted91,92,95,251 UV simple—just via a T-piece for merging the HPLC eZuent with photooxidation16,18,64,90,121,142 or TS-THG.81,83–85 HPLC- the NaBH4 reductant solution (or NaBEt4 reagent for Pb48,81) VGAAS techniques for Hg,104–106,109–114,146,147,252–255 in a simple VG manifold—provided that the determination of Pb,48,80,81,89,252–255 Sb,235,236 Se81,82,84–87,96,99,101,102,256 and only hydride-forming species is targeted:1 arsenite, arsenate, Sn81,247 are also documented.Current HPLC-VGAAS MMA and DMA;70,228–234,248–250 i-SbIII and i-SbV,235,236 and approaches are schematically presented in Fig. 4, while the their methylated monomethylstibonic and dimethylstibinic most convincing application examples are compiled in Table 7. hydride-forming species;1 i-SnII/IV (the oxidation states +II So far, the HPLC-HGAAS technique has reached a certain and +IV cannot be distinguished), MeSn3+, Me2 Sn2+, degree of maturity only for As speciation in some important Me3Sn+, Me4 Sn, EtSn3+, Et2 Sn2+, Et3 Sn+, Et4 Sn247 (also biological and environmental materials with relatively high As BuSn3+ and Bu2Sn2+ with low sensitivity65); selenite; and contents (see also Table 7 and refs. 1 and 36): seafood and tellurite. Hence, post-column decompositions and premarine biota,84,90–92,94,95,229,249,251 urine,64,83,91,231,232 blood reduction may be radically avoided in some speciation assays serum,90,230 sediment142,228,251 and soil extracts248 and some of As and Sb.Species-dependent responses are common with environmental waters.70,91,94,121,232–234,243,248,251 HPLC- this straight approach. HGAAS (QTA) and HPLC-UV-HGAAS (QTA) are now Further complications arise in more comprehensive close to a stage of implementation in routine practice for As speciation tasks relating to non-hydride-forming species of speciation.Some specific characteristics, problems and limi- arsenic (Me4As+,18,64,81,83,85 AB,16,18,64,81,83,90–92,94,95,251 tations of HPLC-VGAAS are addressed in the following AC,18,64,81,83,90,91,94,251 phenylarsonate,16,18 o-arsanilate,16,18 discussion. p-arsenilate,18 arsenosugars, arsenolipids, etc.); selenium (selenate,84,87,96,99,101,102 Me3Se+,81,82,84,85,96,99,101,102 seleno- 6.3.1 Scope, sensitivity and LODs. The technique often lacks methionine,86,87,99,101,102 selenoethionine,84,101,102 selenochosensitivity in analyses of real samples, especially for Hg, Sb line,81,82 selenocystine,84,87,102 methylselenonate,84 selenourea, and Sn, while providing marginal LODs for organoleads etc.); and tin (Bu2Sn2+,34 Bu3Sn+,34 Bu4Sn and other ‘heavy’ (10 mg l-1) and organoselenium species (1–5 mg l-1).In alkyltins).1,63,65 On-line decomposition of these organoelement fact, many published application examples have addressed compounds is best expedited by alkaline peroxodisulfate with spiked sample matrices or simple aqueous solutions or microwave91,92,95,251 or UV irradiation16,18,64,90 for As, yieldsamples with higher analyte levels derived from occupational, ing arsenate; UV photooxidation (for Sn);16,63,65 and micronutritional, therapeutic and environmental monitoring wave-assisted bromination which leaves selenium in its tasks.1,35,36,217 It is fair to admit that some alternative optimum oxidation state, i-SeIV.99,101,102 element-specific detectors such as ICP-MS, HG-ICP- Incorporation of an extra downstream step of pre-reduction MS101,108,118,213–215,235 and HGAFS126,217,218 oVer better rela- is optional for As95,121 and Sb but mandatory for SeVI whenever tive LODs (10–100×), thus ranking the direct HPLC-ICP-MS digestion other than bromination has been applied.96 This coupling as the most powerful and very promising rival increases the complexity of hardware and reaction chemistries technique.and entails further dispersion and loss of resolution, and hence Whereas the HG-VGAAS (QTA) detector by itself is able should be avoided as far as possible. to provide instrumental LODs down to 0.01–0.1 ng, there are Several more notes on the instrumental aspects of this several methodological reasons for impaired procedural LODs hyphenation are worth giving. Most commercial CF- or in HPLC-VGAAS and (furthermore) in HPLC-on-line FI-HGAAS systems—in this case both are operating in the decomposition-VGAAS assays: (i) small injection volumes continuous mode—have so far not been designed with a view (typically 50–100 ml ); (ii) dilution of samples/leachates aimed to coupling with chromatography.Hence software cannot at better handling of matrix eVects and column overload and acquire and process chromatograms. Typical carrier flow rates lifetime; (iii) resorting to HG from oxidised, less sensitive (Fl) are around 4–6 ml min-1 and purge gas flow rates may species such as i-AsV 18,64,90–92,94,228,251 or DMA64,94,229,232,233 be as high as 100–200 ml min-1, whereas much smaller flow so as to avoid an extra step of pre-reduction; (iv) peak rates are optimum in HPLC-HGAAS coupling; cf., Fl of the mobile phase (MP) is around 1 ml min-1 with analytical broadening and noise caused by gaining dispersion during 158 J.Anal. At. Spectrom., 1999, 14, 147–162columns and the Ar flow rate may need to be decreased to CVAAS (QTA) for Bi and Hg; FI-VG-ETAAS with in-atomiser trapping for As, Bi, Ge, Hg, Sb, Se, Sn and Te; 100 ml min-1;63–65 hence scaling down and some adaptation are required.The design and size of the GLS64 may also FI-VGAAS (QTA), CF-VGAAS (QTA), FI-VG-ETAAS and CF-VG-ETAAS for redox speciation of selenite and tellurite; happen not to be optimum in view of compatibility with low Fl, flow pulsations because of low Fl with peristaltic pumps, and a ‘first-order’ speciation of the ‘toxicologically relevant’ As (arsenite+arsenate+MMA+DMA) by FI-HGAAS operating with organic solvents, dead volume, stability against flooding or overwasting, washout characteristics, foam and (QTA) and FI-HG-ETAAS.(b) Availability of automated apparatus and accessories but aerosol formation, compatibility with heating, etc., and hence may also need modification.58,60,64,101 Changes inMP composi- better alternative methods existing or more research and better validation called for: HGAAS for Cd, In, Pb and Tl; tion during gradient elution may lead to a drift in sensitivity and baseline and necessitate a longer conditioning time HG-FAAS for Ge; FI- or CF-HGAAS (QTA) with on-line pre-reduction for As, MMA, DMA, Sb, Se and Te; FI-MWD- between runs; sample throughput rates may drop accordingly by a factor of 5–20 compared with direct FI-HGAAS.HGAAS (QTA) for As, Pb, Se and Sn; redox speciation for AsIII and SbIII by FI-HGAAS and CF-HGAAS; FI-ionexchange- HGAAS for enrichment (As, Se, Sn, etc.) and redox 6.3.3 Mercury speciation.HPLC-VGAAS (QTA) for Hg104–106,109–114,146,147 exhibits some peculiarities: (i) analyte or ‘first-order’ speciation. (c) Future commercialisation expected owing to the levels in most real samples are below the instrumental LODs; hence major steps of pre-column derivatization and enrichment promising performance demonstrated at the research level with custom-built hyphenated systems: FI-UV-VGAAS for As, Hg, with reagents with sulfhydryl groups such as dithiocarbamates, 110–114 L-cysteine104,110,146,147 etc., are applied to the ionic Se and Sn; CVAAS (QTA) with in-line pyrolysis chamber for Hg; VG-CT/GC-AAS (FIT) for As, Hg, Pb, Sb, Se and Sn; i-HgII and organomercury species (MeHg+, EtHg+, PhHg+, MeOHg+, EtOHg+)—mostly on-line in flow systems but also HPLC-VGAAS (QTA) for As, Hg and Sb; HPLC-UVVGAAS (QTA) for As, Hg and Sn; HPLC-MWD-VGAAS supplemented by another oV-line preconcentration step by distillation,114 liquid–liquid extraction111 or sorption on C18 (QTA) for As, Hg, Se and Sn. Whereas the hyphenated VGAAS methods for automated cartridges;106,255 (ii) on-line separation of the complexed organomercury derivatives on C18 RP-HPLC columns is on-line decomposition, pre-reduction, enrichment and nonchromatographic speciation will certainly continue to occupy straightforward but more expensive as MP containing 65% v/v acetonitrile are involved;109–114 (iii) post-column an important niche in routine analytical practice, the implementation of a coupled chromatographic technique with decomposition of organomercurials by UV photooxidation109 –114 or wet chemical oxidation104,105,146 is satisfactory VGAAS detection is likely to meet the challenge by some competitive, more sensitive alternative approaches such as (see Table 3), and so are their derivatization with NaBH4 and on-line thermal decomposition in pyrolysis chambers at 650– VG-GC-MIP-AES, HPLC-VGAFS, HPLC-VG-ICP-MS and, especially, direct coupled HPLC-ICP-MS. 750 °C; (iv) persistent blanks due to trapping Hg0 vapour from the laboratory environment by all reagent solutions, in addition The author strongly believes that further development in this area will depend to a large extent on the future moves of to carry-over and memory eVects originating from the on-line preconcentration module, are observed. instrument manufacturers—both the leading ones and smaller, dynamic companies.It is hoped that they will consider the importance of these challenges and provide the market with 6.3.4 Availability and validation. Hyphenated HPLCVGAAS instrumentation with dedicated software is not com- hyphenated instruments and accessories so as to fertilise the important application fields of trace element determination mercially available, although some modules can be obtained from diVerent manufacturers: hydride generators, mercury- and speciation at mg l-1 and sub-mg l-1 levels.vapour generators, quartz cells, FI accessories, HPLC pumps, The author is grateful to the Royal Society of Chemistry for columns, injectors, HPLC and FI plumbing facilities, inteproviding an RSC Journals Grant for International Authors. grators, software, microwave ovens, UV photolysis units, thermostated baths, etc. The validation of speciation procedures is seriously hindered because of the scarce selection or 8 References lack of appropriate certified reference materials (CRMs). 1 J. De¡ dina and D. L. Tsalev, Hydride Generation Atomic Absorption Spectrometry, Wiley, Chichester, 1995. 2 Trace Element Speciation: Analytical Methods and Problems, ed. 7 Conclusion G. E. Batley, CRC Press, Boca Raton, FL, 1989, pp. 1–350. 3 Environmental Analysis Using Chromatography Interfaced with VGAAS and its modern hyphenated options represent an Atomic Spectroscopy, ed. R. M.Harrison and S. Rapsomanikis, important part of current analytical methodology for trace Ellis Horwood, Chichester, 1989, pp. 1–370. element determination and speciation. Apart from the numer- 4 O. F. X. Donard and R. Ritsema, in Environmental Analysis: ous positive assets of VGAAS, decisive considerations are the Techniques, Applications and Quality Assurance, ed. D. Barcelo, widespread availability of AAS apparatus, aVordable price Elsevier, Amsterdam, 1993, pp. 549–606. and moderate running costs, automation with good sampling 5 R.Ritsema, F. M. Martin and Ph. Quevauviller, in Quality Assurance for Environmental Analysis. Method Evaluation within frequencies and well established methodology. Limitations of the Measurements and Testing Programme (BCR), ed. Ph. some hyphenated VGAAS techniques are the lack of sensitivity Quevauviller, E. A. Maier and B. Griepink, Elsevier, Amsterdam, in some applications, more complex chemistry and optimis- 1995, pp. 489–503. ation involved and relatively narrow linear dynamic range. 6 Element Speciation in Bioinorganic Chemistry, ed. S. Caroli, The state-of-the-art with respect to implementation in routine Wiley, New York, 1996. practice and demands for future research and development 7 Z.-l. Fang, Flow Injection Separation and Preconcentration, VCH, Weinheim, 1993, pp. 1–259. could be categorised as follows. 8 Z.-l. Fang, Flow Injection Atomic Absorption Spectrometry, (a) Availability of automated apparatus and accessories with Wiley, Chichester, 1995, pp. 1–306. unquestioned performance characteristics: FI-HGAAS (QTA) 9 B. Welz and M. Sperling, Atomabsorptionsspektrometrie, 4th or CF-HGAAS (QTA) for As, Bi, Sb, Se, Sn and Te; edn., Wiley-VCH,Weinheim, 1997. FI-CVAAS (QTA) or CF-CVAAS (QTA) with or without 10 D. L. Tsalev, M. Sperling and B. Welz, in 6. Colloquium on-line enrichment by amalgamation for Hg; VGAAS (FIT) Atomspektrometrische Spurenanalytik, ed. B. Welz, Bodenseewerk Perkin-Elmer, U� berlingen, 1991, pp. 349–366. with in situ ethylation for Hg, Pb, Se and Sn; FI-MWDJ. Anal. At. Spectrom., 1999, 14, 147–162 15911 B. Welz, D. L. Tsalev and M. Sperling, Anal. Chim. Acta, 1991, University of Leipzig and UFZ–Umweltforschungszentrum Leipzig-Halle, Leipzig-Halle, 1993, pp. 989–1004. 261, 91. 53 B. Welz, Y. He and M. Sperling, Talanta, 1993, 40, 1917. 12 D. L. Tsalev, M. Sperling and B. Welz, Analyst, 1992, 117, 1729. 54 T. Guo and J.Baasner, Talanta, 1993, 40, 1927. 13 D. L. Tsalev, M. Sperling and B. Welz, Analyst, 1992, 117, 1735. 55 T. Guo and J. Baasner, Anal. Chim. Acta, 1993, 278, 189. 14 H. Agemian and A. S. Y. Chau, Anal. Chem., 1978, 50, 13. 56 T. Guo and J. Baasner, J. Autom. Chem., 1996, 18, 217. 15 H. Agemian and J. A. DaSilva, Anal. Chim. Acta, 1979, 104, 285. 57 G. Schlemmer, J. Baasner, T. Guo and W. Erler, poster presented 16 R. H. Atallah and D. A. Kalman, Talanta, 1991, 38, 167. at ANAKON’97, 6–8 April 1997, Konstanz, Abstracts, p. 133. 17 R. H. Atallah and D. A. Kalman, J. Anal. Toxicol., 1993, 17, 87. 58 X.-c. Le, W. R. Cullen and K. J. Reimer, Appl. Organomet. 18 A. G. Howard and L. E. Hunt, Anal. Chem., 1993, 65, 2995. Chem., 1992, 6, 161. 19 H. Matusiewicz and R. E. Sturgeon, Spectrochim. Acta, Part B, 59 X.-c. Le, W. R. Cullen and K. J. Reimer, Talanta, 1993, 40, 185. 1996, 51, 377. 60 X.-c. Le, W. R. Cullen and K. J. Reimer, Talanta, 1994, 41, 495. 20 G.Drasch, L. v. Meyer and G. Kauert, Fresenius’ Z. Anal. 61 A. Meyer and G. Schwedt, LaborPraxis, 1993, 17(4), 44. Chem., 1980, 304, 141. 62 C. P. Hanna and S. A. McIntosh, At. Spectrosc., 1995, 16, 106. 21 R. E. Sturgeon, S. N. Willie and S. S. Berman, Anal. Chem., 63 D. L. Tsalev, M. Sperling and B.Welz, in CANAS’97 Colloquium 1985, 57, 2311. Analytische Atomspektroskopie, ed. G. Werner, University of 22 R. E. Sturgeon, S. N. Willie and S. S. Berman, J. Anal. At. Leipzig and UFZ-Umweltforschungszentrum Leipzig-Halle, Spectrom., 1986, 1, 115.Leipzig-Halle, in the press. 23 R. E. Sturgeon, S. N. Willie and S. S. Berman, Anal. Chem., 64 D. L. Tsalev, M. Sperling and B. Welz, Analyst, 1998, 123, 1703. 1987, 59, 2441. 65 D. L. Tsalev, M. Sperling and B. Welz, Spectrochim. Acta, Part 24 R. E. Sturgeon, S. N. Willie, G. I. Sproule and S. S. Berman, B, submitted for publication. J. Anal. At. Spectrom., 1987, 2, 719. 66 The FIAS-Furnace Technique: Recommended Analytical 25 R.E. Sturgeon, S. N. Willie and S. S. Berman, Anal. Chem., Conditions and General Information, Publication B3212.20, Part 1989, 61, 1867. No. 0993–5204, Bodenseewerk Perkin-Elmer, U� berlingen, 1993. 26 R. E. Sturgeon, S. N. Willie and S. S. Berman, J. Anal. At. 67 S. Tesfalidet and K. Irgum, Anal. Chem., 1989, 61, 2079. Spectrom., 1989, 4, 443. 68 P. E. Carrero and J. F. Tyson, Analyst, 1997, 122, 915. 27 R. E. Sturgeon, S. N. Willie, G. I. Sproule, P. T.Robinson and 69 A. Larraya, M. G. Cobo-Fernandez, M. A. Palacios and S. S. Berman, Spectrochim. Acta, Part B, 1989, 44, 667. C. Camara, Fresenius’ J. Anal. Chem., 1994, 350, 667. 28 H. W. Sinemus, J. Kleiner, B.ziuk and H.-H. Stabel, in 70 M. Gomez, C. Camara, M. A. Palacios and A. Lopez-Gonzalvez, CANAS’93 Colloquium Analytische Atomspektroskopie, ed. Fresenius’ J. Anal. Chem., 1997, 357, 844. K. Dittrich and B. Welz, University of Leipzig and 71 J. L. Burguera, M. Burguera and C.Rivas, Quim. Anal., 1997, UFZ–Umweltforschungszentirum Leipzig-Halle, Leipzig-Halle, 16, 165. 1993, pp. 657–664. 72 S. G. OZey, N. J. Seare, J. F. Tyson and H. A. B. Kibble, Anal. 29 H. W. Sinemus, H.-H. Stabel, B. Radziuk and J. Kleiner, Proc., 1991, 28, 18. Spectrochim. Acta, Part B, 1993, 48, 643. 73 S. G. OZey, N. J. Seare, J. F. Tyson and H. A. B. Kibble, J. Anal. 30 H. W. Sinemus, H.-H. Stabel, B. Radziuk and J. Kleiner, At. Spectrom., 1991, 6, 133. Spectrochim.Acta, Part B, 1993, 48, 1719. 74 J. F. Tyson, S. G. OZey, N. J. Seare, H. A. B. Kibble and 31 J.C. Van Loon, B. Radziuk, N. Kahn, J. Lichwa, F. J. Fernandez C. Fellows, J. Anal. At. Spectrom., 1992, 7, 315. and J. D. Kerber, At. Absorpt. Newsl., 1977, 16, 79. 75 G.-b. Jiang, Z.-m. Ni, S.-r. Wang and H.-b. Han, Fresenius’ 32 M. O. Andreae, Anal. Chem., 1977, 49, 820. J. Anal. Chem., 1989, 334, 27. 33 M. O. Andreae and J. T. Byrd, Anal. Chim. Acta, 1984, 156, 147. 76 H. Emteborg, H.-W.Sinemus, B. Radziuk, D. C. Baxter and 34 L. Ebdon, S. J. Hill and P. Jones, Talanta, 1991, 38, 607. W. Frech, Spectrochim. Acta, Part B, 1996, 51, 829. 35 D. L. Tsalev, Atomic Absorption Spectrometry in Occupational 77 R.-D.Wilken, Fresenius’ J. Anal. Chem., 1992, 342, 795. and Environmental Health Practice, Vol. II: Determination of 78 R. Puk and J. H. Weber, Anal. Chim. Acta, 1994, 292, 175. Individual Elements, CRC Press, Boca Raton, FL, 1984. 79 J.-S. Blais, G.-M.Momplaisir and W. D. Marshall, Anal. Chem., 36 D. L. Tsalev, Atomic Absorption Spectrometry in Occupational 1990, 62, 1161. and Environmental Health Practice, Vol. III: Progress in 80 J. S. Blais and W. D. Marshall, J. Anal. At. Spectrom., 1989, Analytical Methodology, CRC Press, Boca Raton, FL, 1995. 4, 271. 37 X.-p. Yan and Z.-m. Ni, Anal. Chim. Acta, 1994, 291, 89. 81 W. D. Marshall, J. S. Blais and F. C. Adams, in Metal Speciation 38 T. Nakahara, in Advances in Atomic Spectroscopy, ed.in the Environment, NATO ASI Series, vol. G23, ed. J. Sneddon, JAI Press, Greenwich, CT, 1995, vol. 2, pp. 139–178. J. A. C. Broekaert, S. Gu� cer and F. Adams, Springer, Berlin, 39 J. De¡dina, in Encyclopedia of Environmental Analysis and 1990, pp. 253–273. Remediation, ed. R. A. Meyers, Wiley, New York, 1998, vol. 3, 82 J.-S. Blais, A. Huyghues-Despointes, G. M. Momplaisir and pp. 524–542. W. D. Marshall, J. Anal. At. Spectrom., 1991, 6, 225. 40 S. Rapsomanikis, Analyst, 1994, 119, 1429. 83 G. M. Momplaisir, J.-S. Blais, M. Quinteiro and W. D. Marshall, 41 F. E. Smith and E. A. Arsenault, Talanta, 1996, 43, 1207. J. Agric. Food. Chem., 1991, 39, 1448. 42 Microwave Enhanced Chemistry. Fundamentals, Sample 84 G.-M. Momplaisir, T. Lei and W. D. Marshall, Anal. Chem., Preparation and Applications, ed. H. M. Kingston and 1994, 66, 3533. S. J. Haswell, American Chemical Society Professional Reference 85 A. Huyghues-Despointes, G. M. Momplaisir, J.S. Blais and W. Series 274, American Chemical Society, Washington, DC, 1997. D. Marshall, Chromatographia, 1991, 31, 481. 43 R. Rubio, J. Alberti, A. Padro and G. Rauret, Trends Anal. 86 G. Matni, R. Azani, M. R. Van Calsteren, M. C. Bissonnette and Chem., 1995, 14, 274. J. S. Blais, Analyst, 1995, 120, 395. 44 J. Golimowski and K. Golimowska, Anal. Chim. Acta, 1996, 87 T. Lei and W. D. Marshall, Appl. Organomet. Chem., 1995, 9, 325, 111. 149. 45 S. Rapsomanikis, O. F.X. Donard and J. H. Weber, Anal. 88 Y. Tan, G.-M.Momplaisir, J.Wang and W. D. Marshall, J. Anal. Chem., 1986, 58, 35. At. Spectrom., 1994, 9, 1153. 46 P. J. Craig, D. Mennie, N. Ostah, O. F. X. Donard and 89 C. Tausch, D. Behrendt and B. Neidhart, in CANAS’93 F. Martin, Analyst, 1992, 117, 823. Colloquium Analytische Atomspektroskopie, ed. K. Dittrich 47 R. Fischer, S. Rapsomanikis and M. O. Andreae, Anal. Chem., and B. Welz, University of Leipzig and UFZ– 1993, 65, 763. Umweltforschungszentrum Leipzig-Halle, Leipzig-Halle, 1993, 48 J.S. Blais and W. D. Marshall, J. Anal. At. Spectrom., 1989, pp. 671–675. 4, 641. 90 X. Zhang, R. Cornelis, J. De Kimpe and L. Mees, Anal. Chim. 49 R. Allabashi, J. Rendl and M. Grasserbauer, Fresenius’ J. Anal. Acta, 1996, 319, 177. Chem., 1998, 360, 723. 91 M. A. Palacios, M. Go�mez, C. Ca�mara and M. A. Lo� pez, Anal. 50 S. Shawky, H. Emons and H. W. Du�rbeck, Anal. Commun., 1996, Chim. Acta, 1997, 340, 209. 33, 107. 92 D. Velez, N. Ybanez and R. Montoro, J. Anal. At. Spectrom., 51 J. Szpunar, M. Ceulemans, V. O. Schmitt, F. C. Adams and 1997, 12, 91. R. Lobinski, Anal. Chim. Acta, 1996, 332, 225. 93 S. N. Willie, Spectrochim. Acta, Part B, 1996, 51, 1781. 52 M. Sperling, Y. He and B. Welz, in CANAS’93 Colloquium 94 M. A. Lo�pez, M. M. Go�mez, M. A. Palacios and C. Ca� mara, Fresenius’ J. Anal. Chem., 1993, 346, 643. Analytische Atomspektroskopie, ed. K. Dittrich and B. Welz, 160 J. Anal. At.Spectrom., 1999, 14, 147–16295 K. J. Lamble and S. J. Hill, Anal. Chim. Acta, 1996, 334, 261. 138 A. D’Ulivo, L. Lampugnani, I. Sfetsios, R. Zamboni and C. Forte, Analyst, 1994, 119, 633. 96 M. G. Cobo-Fernandez, M. A. Palacios, D. Chakraborti, 139 I. D. Brindle and E. Lugowska, Spectrochim. Acta, Part B, 1997, P. Quevauviller and C. Camara, Fresenius’ J. Anal. Chem., 1995, 52, 163. 351, 438. 140 S. J. Hill, L. Pitts and P. Worsfold, J. Anal. At. Spectrom., 1995, 97 E.Munaf, H. Haraguchi and T. Takeuchi, Anal. Chim. Acta, 10, 409. 1991, 243, 247. 141 C. Brunori, M. B. de la Calle-Guntinas and R. Morabito, 98 C. P. Hanna, J. F. Tyson and S. McIntosh, Anal. Chem., 1993, Fresenius’ J. Anal. Chem., 1998, 360, 26. 65, 653. 142 H. Narasaki, Y. Kato and H. Kimura, Anal. Sci., 1992, 8, 893. 99 N. Ellend, C. Rohrer, M. Grasserbauer and J. A. C. Broekaert, 143 O. Jimenez de Blas, S. V. Gonzalez, R. Seisdedos Rodriguez and Fresenius’ J. Anal. Chem., 1996, 356, 99.J. Hernandez Mendez, J. Assoc. OV. Anal. Chem., 1994, 77, 441. 100 J. M. Gonzalez LaFuente, M. L. Fernandez Sanchez, 144 S.-c. Zhang, S.-k. Xu and Z.-l. Fang, Quim. Anal., 1989, 8, 191. J. M. Marchante-Gayon, J. E. Sanchez Uria and A. Sanz-Medel, 145 S.-c. Zhang, S.-k. Xu and Z.-l. Fang, Guangpuxue Yu Guangpu Spectrochim. Acta, Part B, 1996, 51, 1849. Fenxi, 1988, 8(6), 39. 101 J. M. Gonzalez LaFuente, M. L. Fernandez Sanchez and 146 E. Munaf, H. Haraguchi, D. Ishii, T.Takeuchi and M. Goto, A. Sanz-Medel, J. Anal. At. Spectrom., 1996, 11, 1163. Anal. Chim. Acta, 1990, 235, 399. 102 J. M. Marchante-Gayon, J. M. Gonzalez, M. L. Fernandez, 147 C. Sarzanini, G. Sacchero, M. Aceto, O. Abollino and E. Blanco and A. Sanz-Medel, Fresenius’ J. Anal. Chem., 1996, E. Mentasti, Anal. Chim. Acta, 1994, 284, 661. 355, 615. 148 S.-k. Xu and Z.-l. Fang, Chin. Chem. Lett., 1992, 3, 915. 103 T. Guo, J. Baasner, M. Gradl and A. Kistner, Anal. Chim. Acta, 149 J.-y.Cao and H. Narasaki, Bunseki Kagaku, 1994, 43, 169. 1996, 320, 171. 150 S.-k. Xu, S.-c. Zhang and Z.-l. Fang, Kexue Tongbao, 1989, 34, 104 F. Palmisano, P. G. Zambonin and N. Cardellicchio, Fresenius’ 1359. J. Anal. Chem., 1993, 346, 648. 151 Z.-l. Fang, L.-j. Sun, E. H. Hansen, J. E. Olesen and 105 C. Schickling and J. A. C. Broekaert, in CANAS’95 Colloquium L. M. Henriksen, Talanta, 1992, 39, 383. Analytische Atomspektroskopie, ed. B. Welz, Bodenseewerk 152 G.Tao and Z. Fang, Talanta, 1995, 42, 375. Perkin-Elmer, U� berlingen, 1996, pp. 407–412. 153 Y. An, S. N. Willie and R. E. Sturgeon, Spectrochim. Acta, Part 106 J. C. G. Wu, Spectrosc. Lett., 1991, 24, 681. B, 1992, 47, 1403. 107 R. Rubio, J. Alberti and G. Rauret, Int. J. Environ. Anal. Chem., 154 H.-b. Han, Y.-b. Liu, S.-f. Mou and Z.-m. Ni, J. Anal. At. 1993, 52, 203. Spectrom., 1993, 8, 1085. 108 J. M. Gonzalez LaFuente, M. Dlaska, M. L. Fernandez Sanchez 155 H. Goenaga Infante, M.L. Fernandez Sanchez and A. Sanz- and A. Sanz-Medel, J. Anal. At.13, 423. Medel, J. Anal. At. Spectrom., 1996, 11, 571. 109 R. Falter and H. F. Scho�ler, Fresenius’ J. Anal. Chem., 1994, 156 P. Bermejo-Barrera, J. Moreda-Pineiro, A. Moreda-Pineiro and 348, 253. A. Bermejo-Barrera, J. Anal. At. Spectrom., 1996, 11, 1081. 110 R. Falter and H. F. Scho� ler, J. Chromatogr. A, 1994, 675, 253. 157 H. Matusiewicz, M. Kopras and R. E. Sturgeon, Analyst, 1997, 111 R.Falter and H. F. Scho� ler, Chemosphere, 1994, 29, 1333. 122, 331. 112 R. Falter and H. F. Scho�ler, Fresenius’ J. Anal. Chem., 1995, 158 Z.-m. Ni and D.-q. Zhang, Spectrochim. Acta, Part B, 1995, 353, 34. 50, 1779. 113 R. Falter and H. F. Scho�ler, Fresenius’ J. Anal. Chem., 1996, 159 Z.-m. Ni and B. He, J. Anal. At. Spectrom., 1995, 10, 747. 354, 492. 160 G. Tao and Z. Fang, J. Anal. At. Spectrom., 1993, 8, 577. 114 R. Eiden, R. Falter, B. Augustin-Castro and H. F.Scho� ler, 161 H. O. Haug and C. Ju, J. Anal. At. Spectrom., 1990, 5, 215. Fresenius’ J. Anal. Chem., 1997, 357, 439. 162 H. O. Haug and Y. Liao, J. Anal. At. Spectrom., 1995, 10, 1069. 115 A. D’Ulivo, Analyst, 1997, 122, 117R. 163 B. Hilligsoe, J. E. T. Andersen and E. H. Hansen, J. Anal. At. 116 W. Driehaus and M. Jekel, Fresenius’ J. Anal. Chem., 1992, 343, Spectrom., 1997, 12, 585. 352. 164 D.-q. Zhang and Z.-m. Ni, Anal. Chim. Acta, 1996, 330, 53. 117 C. P. Hanna, J. F.Tyson and S. McIntosh, Clin. Chem., 1993, 165 X.-p. Yan, Z.-m. Ni and Q.-l. Guo, Anal. Chim. Acta, 1993, 39, 1662. 272, 105. 118 C.-j. Hwang and S.- j. Jiang, Anal. Chim. Acta, 1994, 289, 205. 166 P. Bermejo-Barrera, J. Moreda-Pineiro, A. Moreda-Pineiro and 119 S. Nielsen and E. H. Hansen, Anal. Chim. Acta, 1997, 343, 5. A. Bermejo-Barrera, J. Anal. At. Spectrom., 1997, 12, 317. 120 D. Schaumlo�Vel and B. Neidhart, Fresenius’ J. Anal. Chem., 167 Y.-p. Liao and A.-m. Li, J. Anal.At. Spectrom., 1993, 8, 633. 1996, 354, 866. 168 X.-p. Yan and Z.-m. Ni, J. Anal. At. Spectrom., 1991, 6, 483. 121 J. Stummeyer, B. Harazim and T. Wippermann, T., Fresenius’ 169 H. O. Haug, Spectrochim. Acta, Part B, 1996, 51, 1425. J. Anal. Chem., 1996, 354, 344. 170 S. N. Willie, At. Spectrosc., 1994, 15, 205. 122 D. L. Tsalev, M. Sperling and B. Welz, Talanta, submitted for 171 D. Erber, J. Bettmer and C. Cammann, Fresenius’ Z. Anal. publication. Chem., 1994, 349, 738. 123 F. Nakata, H. Sunahara, H. Matsuo and T. Kumamaru, Bunseki 172 A. D’Ulivo, L. Lampugnani, D. Faraci, D. L. Tsalev and Kagaku, 1986, 35, 439. R. Zamboni, Talanta, 1998, 44, 801. 124 M. G. Cobo Fernandez, M. A. Palacios and C. Camara, Anal. 173 A. M. Wifladt, G. Wibetoe and W. Lund, Fresenius’ J. Anal. Chim. Acta, 1993, 283, 386. Chem., 1997, 357, 92. 125 J. L. Burguera, P. Carrero, M. Burguera, C. Rondon, 174 H.-W. Sinemus, J. Kleiner, H.-H. Stabel and B. Radziuk, J. Anal. M. R.Brunetto and M. Callignani, Spectrochim. Acta, Part B, At. Spectrom., 1992, 7, 433. 1996, 51, 1837. 175 G. Schlemmer and M. Feuerstein, in CANAS’93 Colloquium 126 L. Pitts, A. Fisher, P. Worsfold and S. J. Hill, J. Anal. At. Analytische Atomspektroskopie, ed. K. Dittrich and B. Welz, Spectrom., 1995, 10, 519. University of Leipzig and UFZ–Umweltforschungszentrum 127 K. Jin, M. Taga, H. Joshida and S. Hikime, Bull. Chem. Soc. Leipzig-Halle, Leipzig-Halle, 1993, pp. 431–441. Jpn., 1979, 52, 2276. 176 C. P. Hanna, G. R. Carnrick, S. A. McIntosh, L. C. Guyette and 128 A. G. Howard and C. Salou, Anal. Chim. Acta, 1996, 333, 89. D. E. Bergemann, At. Spectrosc., 1995, 16, 82. 129 X.-C. Le, W. R. Cullen and K. J. Reimer, Anal. Chim. Acta, 177 D.-q. Zhang, H.-w. Sun and L-l. Yang, Fresenius’ J. Anal. Chem., 1994, 285, 277. 1997, 359, 492. 130 A. G. Howard, J. Anal. At. Spectrom., 1997, 12, 267. 178 L. Zhang, Z.-m. Ni and X.-q. Shan, Can. J. Appl. Spectrosc., 131 D.L. Tsalev, A. D’Ulivo, L. Lampugnani, M. Di Marco and 1991, 36, 47. R. Zamboni, J. Anal. At. Spectrom., 1996, 11, 989. 179 G.-b. Jiang, Z.-m. Ni, L. Zhang, A. Li, H.-b. Han and 132 T. Guo, J. Baasner and D. L. Tsalev, Anal. Chim. Acta, 1997, X.-q. Shan, J. Anal. At. Spectrom., 1992, 7, 447. 349, 313. 180 Z.-m. Ni, B. He and H.-b. Han, Can. J. Appl. Spectrosc., 1993, 133 I. D. Brindle and C. M. Ceccarelli Ponzoni, Analyst, 1987, 112, 38, 11. 1547. 181 J. F. Tyson, N. G. Sundin, C.P. Hanna and S. A. McIntosh, 134 B. Welz and M. S¡ ucmanova�, Analyst, 1993, 118, 1417. Spectrochim. Acta, Part B, 1997, 52, 1773. 135 B. Welz and M. S¡ ucmanova�, Analyst, 1993, 118, 1425. 182 J. Y. Cabon and W. Erler, Analyst, 1998, 123, 1565. 136 A. D’Ulivo, J. Anal. At. Spectrom., 1989, 4, 67. 183 D. Erber, W. Buscher and K. Cammann, in CANAS’95 137 A. D’Ulivo, L. Lampugnani, I. Sfetsios and R. Zamboni, Colloquium Analytische Atomspektroskopie, ed. B. Welz, Bodenseewerk Perkin-Elmer, U� berlingen, 1996, pp. 321–328. Spectrochim. Acta, Part B, 1993, 48, 387. J. Anal. At. Spectrom., 1999, 14, 147–162 161184 H. G. Riepe, D. Erber, J. Bettmer and K. Cammann, Fresenius’ 221 P. Quevauviller, F. Martin, C. Belin and O. F. X. Donard, Appl. Organomet. Chem., 1993, 7, 149. J. Anal. Chem., 1997, 359, 239. 222 F. M. Martin, C.-M. Tseng, C. Belin, P. Quevauviller and 185 Z.-m. Ni, H.-b. Hang, A. Li, B. He and F.-z. Xu, J. Anal. At. O. F. X. Donard, Anal.Chim. Acta, 1994, 286, 343. Spectrom., 1991, 6, 385. 223 F. Pannier, A. Astruc and M. Astruc, Anal. Chim. Acta, 1996, 186 H. O. Haug and Y. Liao, Spectrochim. Acta, Part B, 1995, 50, 327, 287. 1311. 224 M. Astruc, R. Pinel and A. Astruc, Mikrochim. Acta, 1992, 109, 187 L. Zhang, S. McIntosh, G. R. Carnrick and W. Slavin, 73. Spectrochim. Acta, Part B, 1992, 47, 701. 225 P. M. Sarradin, F. Leguille, A. Astruc, R. Pinel and M. Astruc, 188 B. M. Yoon, S. C. Shim, H. C. Pyun and D.S. Lee, Anal. Sci., Analyst, 1995, 120, 79. 1990, 6, 561. 226 D. Dyne, B. S. Chana, N. J. Smith and J. Cocker, Anal. Chim. 189 M. Walcerz, S. Garbos, E. Bulska and A. Hulanicki, Fresenius’ Acta, 1991, 246, 351. J. Anal. Chem., 1994, 350, 662. 227 W. M. R. Dirkx, W. E. Van Mol, R. J. A. Van Cleuvenbergen 190 R. Kala�hne, G. Henrion, A. Hulanicki, S. Garbos and and F. C. Adams, Fresenius’ J. Anal. Chem., 1989, 335, 769. M.Walcerz, Spectrochim. Acta, Part B, 1997, 52, 1509. 228 B. A. Manning and D. A. Martens, Environ. Sci. Technol., 1997, 191 M. M. Chaudhry, A. M. Ure, B. Cooksey, D. Littlejohn and 31, 171. D. J. Halls, Anal. Proc., 1991, 28, 44. 229 D. Velez, N. Ybanez and R. Montoro, J. Anal. At. Spectrom., 192 S. Garbos, M. Walcerz, E. Bulska and A. Hulanicki, 1996, 11, 271. Spectrochim. Acta, Part B, 1995, 50, 1669. 230 X. Zhang, R. Cornelis, J. De Kimpe and L. Mees, J. Anal. At. 193 M. Veber, K. Cujes and S. Gomiscek, J. Anal. At. Spectrom., Spectrom., 1996, 11, 1075. 1994, 9, 285. 231 B. S. Chana and N. J. Smith, Anal. Chim. Acta, 1987, 197, 177. 194 W.-W. Ding and R. E. Sturgeon, Spectrochim. Acta, Part B, 232 E. Hakala and L. Pyy, J. Anal. At. Spectrom., 1992, 7, 191. 1996, 51, 1325. 233 F.-h. Ko, S.-l. Chen and M.-h. Yang, J. Anal. At. Spectrom., 195 J. Murphy, P. Jones, S. J. Hill, G. Schlemmer and I. Shuttler, 1997, 12, 589. Anal. Commun., 1997, 34, 359. 234 J. Stummeyer, B. Harazim and T. Wippermann, in CANAS’95 196 Z.-m.Ni, B. He and H.-b. Han, J. Anal. At. Spectrom., 1993, Colloquium Analytische Atomspektroskopie, ed. B. Welz, 8, 995. Bodenseewerk Perkin-Elmer, U� berlingen, 1996, pp. 425–428. 197 Y.-p. Liao and H. O. Haug, Microchem. J., 1997, 56, 247. 235 P. Smichowski, Y. Madrid, M. B. de la Calle Guntinas and 198 H. O. Haug and Y.-p. Liao, Fresenius’ J. Anal. Chem., 1996, C. Camara, J. Anal. At. Spectrom., 1995, 10, 815. 356, 435. 236 X. Zhang, R. Cornelis and L.Mees, J. Anal. At. Spectrom., 1998, 199 I. Shuttler, M. Feuerstein and G. Schlemmer, J. Anal. At. 13, 205. Spectrom., 1992, 7, 1299. 237 T. R. Ru�de and H. Puchelt, Fresenius’ J. Anal. Chem., 1994, 200 D. L. Tsalev, A. D’Ulivo, L. Lampugnani, M. Di Marco and 350, 44. R. Zamboni, J. Anal. At. Spectrom., 1995, 10, 1003. 238 X. Yin, E. HoVman and C. Lu�dke, Fresenius’ J. An A. D’Ulivo, L. Lampugnani, M. Di Marco and 1996, 355, 324. R. Zamboni, J. Anal. At. Spectrom., 1996, 11, 979. 239 Y. Cai, S. Rapsomanikis and M. O. Andreae, Mikrochim. Acta, 202 P. S. Doidge, B. T. Sturman and T. M. Rettberg, J. Anal. At. 1992, 109, 67. Spectrom., 1989, 4, 251. 240 M. Ceulemans, J. Szpunar-Lobinska, W. M. R. Dirkx, 203 Y. An, S.N.Willie and R. E. Sturgeon, Fresenius’ J. Anal. Chem., R. Lobinski and F. C. Adams, Int. J. Environ. Anal. Chem., 1993, 1992, 344, 64. 5, 113. 204 L. Zhang, Z.-m. Ni and X.-q. Shan, Spectrochim. Acta, Part B, 241 L. Liang, M. Horvat, E. Cernichiari, B. Gelein and S. Balogh, 1989, 44, 339. Talanta, 1996, 43, 1883. 205 L. Zhang, Z.-m. Ni and X.-q. Shan, Spectrochim. Acta, Part B, 242 I. R. Pereiro, V. O. Schmitt, J. Szpunar, O. F. X. Donard and 1989, 44, 751. R. Lobinski, Anal. Chem., 1996, 68, 4135. 206 J. De¡dina, T. Matousek and W. Frech, Spectrochim. Acta, Part 243 L. Ebdon, S. Hill, A. P. Walton and R. W. Ward, Analyst, 1988, B, 1996, 51, 1107. 113, 1159. 207 W. Slavin, D. C. Manning and G. R. Carnrick, At. Spectrosc., 244 K. J. Reimer and J. A. J. Thompson, Biogeochemistry, 1988, 1981, 2, 137. 6, 211. 208 B. Docekal, J. De¡dina and V. Krivan, Spectrochim. Acta, Part B, 245 R. Kruse, U. Ballin and H.-A. Ru� ssel-Sinn, in CANAS’93 1997, 52, 787. Colloquium Analytische Atomspektroskopie, ed. K. Dittrich 209 V. I. Slaveykova, L. Lampugnani, D. L. Tsalev and L. Sabbatini, and B. Welz, University of Leipzig and UFZ– Spectrochim. Acta, Part B, 1997, 52, 2115. Umweltforschungszentrum Leipzig-Halle, Leipzig-Halle, 1993, 210 V. I. Slaveykova, L. Lampugnani, D. L. Tsalev, L. Sabbatini and pp. 557–564. E. De Giglio, Spectrochim. Acta, Part B, in the press. 246 U. Ballin, R. Kruse and H.-A. Ru� ssel, Fresenius’ J. Anal. Chem., 211 D. L. Tsalev, V. I. Slaveykova and P. B. Mandjukov, 1994, 350, 54. Spectrochim. Acta Rev., 1990, 13, 225. 247 D. T. Burns, F. Glockling and M. Harriot, Analyst, 1981, 106, 212 D. L. Tsalev and V. I. Slaveykova, in Advances in Atomic 921. Spectroscopy, ed. J. Sneddon, JAI Press, Greenwich, CT, 1998, 248 C. T. Tye, S. J. Haswell, P. O’Neill and K. C. C. Bancroft, Anal. vol. 4, ch. 2, pp. 27–150. Chim. Acta, 1985, 169, 195. 213 D. T. Heitkemper, K. A. Wolnik, F. L. Fricke and J. A. Caruso, 249 T. Maitani, S. Uchiyama and Y. Saito, J. Chromatogr., 1987, in Inductively Coupled Plasmas in Analytical Atomic 391, 161. Spectrometry, ed. A. Montaser and D. W. Golighty, VCH, 250 S. Branch, K. C. C. Bancroft, L. Ebdon and P. O’Neill, Anal. Weinheim, 2nd edn., 1992, ch. 17, pp. 781–834. Proc., 1989, 26, 73. 214 J. M. Carey, F. A. Byrdy and J. A. Caruso, J. Chromatogr. Sci., 251 M. A. Lo�pez-Gonza�lvez, M. M. Go�mez, C. Ca�mara and 1993, 31, 330. M. A. Palacios, J. Anal. At. Spectrom., 1994, 9, 291. 215 K. Sutton, R. M. C. Sutton and J. A. Caruso, J. Chromatogr. A, 252 M. Fujita and E. Takabatake, Anal. Chem., 1983, 55, 454. 1997, 789, 85. 253 W. Holak, J. Liq. Chromatogr., 1985, 8, 563. 216 L. Ebdon, S. Hill and R. W.Ward, Analyst, 1986, 111, 1113. 254 W. Holak, J. Assoc. OV. Anal. Chem., 1989, 72, 926. 217 L. Ebdon, S. Hill and R. W.Ward, Analyst, 1987, 112, 1. 255 B. Aizpun-Fernandez, M. L. Fernandez, E. Blanco and A. Sanz-Medel, J. Anal. At. Spectrom., 1994, 9, 1279. 218 A. D’Ulivo, in Environmental Analysis Using Chromatography 256 A. Meyer and L. Dunemann, in 6. Colloquium Interfaced with Atomic Spectroscopy, ed. R. M. Harrison and Atomspektrometrische Spurenanalytik, ed. B. Welz, S. Rapsomanikis, Ellis Horwood, Chichester, 1989, ch. 5. Bodenseewerk Perkin-Elmer, U� berlingen, 1991, pp. 315–321. 219 E. Bulska, J. Anal. At. Spectrom., 1992, 7, 201. 220 K. Bergmann and B. Neidhart, Fresenius’ J. Anal. Chem., 1996, 356, 57. Paper 8/07304J 162 J. Anal. At. Spectrom., 1999, 14,
ISSN:0267-9477
DOI:10.1039/a807304j
出版商:RSC
年代:1999
数据来源: RSC
|
9. |
Operational speciation of uranium in inter-tidal sediments from the vicinity of a phosphoric acid plant by means of the BCR sequential extraction procedure and ICP-MS |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 163-168
Susan E. Howe,
Preview
|
|
摘要:
Operational speciation of uranium in inter-tidal sediments from the vicinity of a phosphoric acid plant by means of the BCR sequential extraction procedure and ICP-MS† Susan E. Howe,a Christine M. Davidson,*a and Martin McCartneyb aDepartment of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow, UK G1 1XL bScottish Universities Research and Reactor Centre, Rankine Avenue, East Kilbride, UK G75 0LF Received 1st September 1998, Accepted 6th November 1998 A method has been developed for the determination of uranium in sediment digests and sequential extracts by ICP-MS.No interferences were found due to the presence of the extractant matrices (0.11 mol l-1 acetic acid, 0.1 mol l-1 hydroxylammonium chloride, 1.0 mol l-1 ammonium acetate and 30% v/v aqua regia) provided 236U was used as internal standard and the samples were diluted 5-fold with 5% v/v nitric acid prior to analysis. Detection limits (8 ng g-1 for extracts and 20 ng g-1 for digests) were adequate for determination of uranium at environmental levels.Results obtained by ICP-MS for aqua regia digests were slightly lower than those obtained by alpha spectrometry, but overall recoveries of uranium by sequential extraction were generally within ±10% of pseudototal values. Application of the developed method to inter-tidal sediments collected from the vicinity of the Albright and Wilson phosphoric acid plant, Whitehaven, Cumbria, confirmed the presence of significant technological enhancement of uranium at Whitehaven Harbour but showed that concentrations had decreased since the plant discontinued phosphate ore processing in 1992.Results of the sequential extraction suggested that spillage of phosphate ore (containing up to 120 mg g-1 uranium) during shipping may have made a significant contribution to the environmental contamination observed. Chemical speciation is important in order to assess the Less well recognised, however, is the technological enhancepotential of sediments to act as sources or sinks of heavy ment of natural radioactivity by conventional (non-nuclear) metal contaminants.While precise chemical forms of elements industries such as mining and fossil fuel exploitation.22 Of such as mercury and tin can be determined in sediment particular concern is the phosphates industry and technological samples, this is not generally possible for other environmentally enhancement of natural radioactivity due to phosphates proimportant metals.Operational methods of speciation such as duction is widely reported.23–30 The source of this enhancement sequential extraction are therefore used to indicate the fraction is sedimentary phosphate ore, which contains significant conof the total metal content which is potentially labile under centrations of uranium, together with radiologically important various environmental conditions. Many diVerent sequential daughter nuclides such as 226Ra, 210Pb and 210Po.These may extraction schemes have been developed.1 However, recently, be concentrated in products/by-products or released to the in an attempt to harmonise methodology, the Commission of environment during phosphate ore processing.23 the European Communities, Community Bureau of Reference Interest in technological enhancement in the UK has focused (BCR, now the Standards, Measurement and Testing on the Albright and Wilson, ‘Marchon’ phosphoric acid Programme) proposed a simple, three-stage protocol2 and production plant, near Whitehaven, Cumbria, NW prepared a reference sediment certified for trace metals extract- England.28,29 From the early 1950s until 1992, the plant able by the procedure.3 The BCR procedure has been applied imported some 300 000 tonnes of phosphate ore per year, successfully to a variety of sediments,4–8 sewage sludge9 and containing approximately 120 mg g-1 uranium.Each year, soils.10,11 phosphogypsum waste containing around 30 tonnes of uran- Relatively few applications of sequential extraction for the ium was discharged, under authorisation, at Saltom Bay, some operational speciation of radionuclides have been reported.12,13 3 km south of Whitehaven. Recent estimates suggest that this Bunzl and co-workers applied sequential extraction (a modified may have caused significant (up to 6 mSv a-1) exposure of Tessier14 method) to study natural radionuclides in the vicinity local seafood consumers, mainly through the ingestion of of an exhaust shaft from a uranium mine15 and to assess 210Pb and 210Po.31 chemical fractionation in a soil contaminated with slag.16 In 1992, on-site processing of phosphate ore at the Similar methods were used to investigate the associated forms Whitehaven plant ceased in favour of import of a crude grade of anthropogenic (fallout) nuclides.17–19 Shultz et al.evaluated of phosphoric acid.Construction and commissioning of a rigorously a five-step sequential extraction procedure for actin- raYnate treatment plant was also undertaken. These alteride speciation,20 whilst Pulford and co-workers applied a six- ations brought about a marked reduction in the amounts of step protocol to speciate Sellafield-derived radionuclides in radioactivity discharged32 which was reflected in a general salt marsh deposits.21 decrease in levels of both 210Po and 238U in sediment and shell There is widespread public awareness of the current and fish in the vicinity of the plant.30 However, in 1996, significant potential impact of the nuclear industry on the environment.technological enhancement of uranium and its decay products remained at Whitehaven Harbour.33 The exact source of the enhanced levels of natural radionuclides at this location is †Presented at the Ninth Biennial National Atomic Spectroscopy Symposium (BNASS), Bath, UK, July 8–10, 1998. unclear.Spillage has been observed to occur from cargo vessels J. Anal. At. Spectrom., 1999, 14, 163–168 163and barges transporting bulk raw materials into the harbour. south (Nethertown and Ravenglass) were also sampled. Information on the samples is given in Table 2. Inter-tidal It is therefore possible that uranium contamination may have resulted from accidental loss of ore overboard, during transfer sediment was collected using an acid-rinsed plastic scoop and taken to the laboratory in polyethylene bags.The sediment and unloading, as well as discharge of waste phosphogypsum slurry at Saltom Bay. was air dried and large aggregates broken up. Stones and large shell fragments were discarded and the material that passed In the present study, sediment collected over a five-year period from Whitehaven Harbour and from other locations through a 1 mm, nylon mesh was retained for analysis. on the Cumbrian coast was sequentially extracted in order to Microwave assisted digestion.For determination of assess the potential mobility and radiological significance of pseudototal (i.e., aqua regia-extractable) concentrations, 1 g of the uranium present. Samples of phosphate ore and of phosair dried sediment was weighed into a PTFE bomb and 7.5 ml phogypsum waste obtained from the Marchon plant were also HNO3 and 22.5 ml HCl were added. The bomb was sealed studied in order to assess the usefulness of sequential extraction and the sample digested in a laboratory microwave oven for discriminating between potential sources of contamination. (CEM-MDS-2000, CEM Corporation, North Carolina, USA).Uranium was determined by inductively coupled plasma mass After cooling, the contents were filtered (Whatman type 592 spectrometry as this is a well established technique for the filter paper) into a 100 ml volumetric flask and made up to determination of long-lived radionuclides and presents a more the mark with 5% HNO3.rapid alternative to the traditional method of alpha particle spectrometry.34,35 Sequential extraction. The reagents used in the BCR sequential extraction protocol, together with the nominal Experimental sediment phases dissolved at each step, are given in Table 3. The experimental procedure is summarised below and has Instrumentation been described in detail in a previous publication.36 Uranium was determined in digests and extracts by ICP-MS. Step One. A 40 ml volume of 0.11 mol l-1 acetic acid was The instrument used was a VG PlasmaQuad PQ II Turbo Plus added to 1 g dried sediment in a polypropylene centrifuge tube (Fisons Instruments, Winsford, Cheshire, UK) with a nebul- and shaken overnight, at ambient temperature, on a mechanizer (Meinhard, Santa Ana, CA, USA) and Scott-type dual ical, end-over-end shaker.The mixture was centrifuged and pass spray chamber. Operating conditions are given in Table 1. the supernatant retained for analysis.Step Two. A 40 ml volume of 0.1 mol l-1 hydroxylam- Reagents monium chloride (adjusted to pH 2 with nitric acid) was added to the residue from Step One, and the extraction performed Acetic acid, hydroxylammonium chloride, ammonium acetate, as described above. hydrochloric acid and the nitric acid used in microwave Step Three. A 10 ml volume of 8.8 mol l-1 hydrogen assisted digestion were of AnalaR grade (Merck, Lutterworth, peroxide was added to the residue from Step Two.The Leicestershire, UK). Nitric acid used as a diluent and to adjust centrifuge tube was covered and the contents digested for 1 h the pH of the extractants was of Aristar grade, also from at ambient temperature, then for 1 h at 85 °C. The sample was Merck. Hydrogen peroxide was of Puriss grade (Fluka, evaporated to near dryness and the hydrogen peroxide treat- Gillingham, Dorset, UK). ment repeated. After taking to near dryness a second time, A uranium standard solution (‘Specpure’, 1000 mg ml-1) 50 ml of 1 mol l-1 ammonium acetate (adjusted to pH 2 with was obtained from Johnson-Matthey (Herts., UK) and was nitric acid) was added and the extraction performed as diluted volumetrically with 5% (v/v) nitric acid to obtain described above.working standards used in calibration. A 236U solution, of The material remaining after sequential extraction was certified specific activity 1.625 Bq ml-1 (equivalent concendigested in 30 ml aqua regia as described above.Due to limited tration 0.678 mg l-1), was obtained from Isotrak, Amersham, sample, only 0.25 g aliquots of the phosphate ores and phos- Buckinghamshire, UK. Certified reference material CRM 601 phogypsum waste were extracted. Reagent volumes were was obtained from the Institute for Reference Materials, Geel, decreased proportionally, to maintain the same solid : extract- Belgium. All dilutions were performed with distilled water. ant ratio as for the sediment samples.Up to twelve samples could be extracted simultaneously, and the total time to Procedures process a batch of samples (including overnight shaking and Sampling. Sediment samples were collected every 6–12 analysis) was of the order of five days. months from Whitehaven Harbour in the period November 1992 to October 1997. On one occasion (November 1992) one Analysis. Subsamples of the extracts and digests (#1 ml, site to the north of Whitehaven (Maryport) and two to the measured accurately) were spiked with 5 ng 236U and diluted to 5 ml with 5% v/v nitric acid prior to introduction to the ICP-MS.Table 1 Instrumental operating conditions Parameter Results and discussion Rf power/W 1350 Method development and characterisation Argon gas flow rates/l min-1: Outer 14 Intermediate 0.75 Uptake and washout times were investigated by sequentially Aerosol 0.87 analysing a 5% v/v nitric acid blank, a 10 ng ml-1 uranium Sample uptake rate/ml min-1 0.8 standard solution, and a further acid blank.Twenty-five Acquisition time/s 30 replicate measurements were performed on each solution, and Measurements per sample 3 each measurement lasted for 10 s. The time taken to obtain Dwell time/ms 10.24 Collection mode Peak jumping the maximum intensity for the uranium signal after supply of Points per peak 3 the standard, or for the signal to reach minimum intensity Sampling cone Ni, 1 mm orifice after supply of the blank, was 50 s. Conservative uptake and Skimmer cone Ni, 0.75 mm orifice washout times of 90 s and 180 s, respectively, were chosen to Sampling distance 10 mm from load coil ensure freedom from memory eVects. 164 J. Anal. At. Spectrom., 1999, 14, 163–168Table 2 Sample characteristics Particle size distributionc (%) Moisturea Loss on Sample Location Date (%) ignitionb (%) <63 mm 63–250 mm >250 mm WH1 Whitehaven Nov. 1992 1.3 5.3 25 58 17 WH2 Mar. 1993 1.6 10.8 19 53 28 WH3 Aug. 1993 1.8 7.3 22 47 31 WH4 Feb. 1994 0.8 2.7 3 82 15 WH5 Jul. 1994 2.6 10.9 35 42 23 WH6 Dec. 1994 1.1 5.0 9 77 14 WH7 Nov. 1996 1.2 6.4 15 61 24 WH8 Oct. 1997 1.2 4.2 7 57 36 MY1 Maryportd Nov. 1992 1.2 3.3 5 86 9 NT1 Nethertowne Nov. 1992 1.4 2.5 5 21 74 RG1 Ravenglassf Nov. 1992 1.1 1.0 7 59 34 CRM 3.4 12.8 Ore K 2.8 NDg Ore Y 2.5 ND PPGh 20.7 ND aMoisture content determined by drying to constant weight at 100°C. bLoss on ignition determined by ashing to constant weight at 550 °C. cFor the <1 mm fraction.d19 km north of Whitehaven. e11 km south of Whitehaven. f21 km south of Whitehaven. gND=not determined. hPPG=phosphogypsum. Table 3 BCR three-stage sequential extraction scheme and with the aqua regia digest were more severe, however, and a decrease in precision was noted when either 115In or 209Bi Nominal target was used. The performance of 236U was consistently good Reagent(s) phase(s) (RSD <1%, for n=5, in all matrices). This is as expected since the standard is very similar in mass to the analyte and Step 1 CH3COOH Exchangeable, water and acid soluble has the same ionisation characteristics in the plasma.The Step 2 NH2OH·HCl at pH 2 Reducible (e.g., Fe/Mn uranium isotope was therefore adopted for use as internal oxyhydroxides) standard in subsequent analyses. It should be noted that 236U Step 3 H2O2, then CH3COONH4 Oxidisable (e.g., organic is a synthetic nuclide and so unlikely to be present in these at pH 2 matter and sulfides) natural samples at a significant level.The low isotopic abun- Residuala Aqua regia dance of 236U, together with the low degree of formation of aNot part of the original BCR protocol. UH+ polyatomic ions in the plasma (generally less than 0.01%)42,43 reduces this potential interference to insignificant levels. A variety of internal standards have been used in the The long term precision and stability of the method were determination of uranium by ICP-MS, including isotopes of demonstrated by storing and repeatedly analysing a set of indium,37 iridium,38 thallium37,39 and bismuth.40 Uranium has extracts over a period of six months.Extracts were stored in been determined in extracts obtained by the BCR procedure PTFE bottles at 4 °C, without additional acidification. The using 205Tl.8 However, it is desirable that an internal standard RSD values obtained for analysis of the same set of extracts be as similar as possible to the analyte. 236U, which has been on twenty two separate occasions were: Step 1, 6.2%; Step 2, used successfully in the determination of uranium in corals by 3.7%; Step 3, 5.4%; Residual, 4.9%.ID-ICP-MS,41 was therefore considered. Potential interference eVects associated with the A set of sequential extracts of a sediment was spiked with determination of uranium in digests and extracts were investi- 115In, 209Bi and 236U, so that their performance as internal gated by comparing the response curve in 5% v/v nitric acid standards for the determination of uranium could be comwith standard addition graphs obtained in acetic acid, pared.The nuclides were added (together) to 5 ml aliquots of hydroxylammonium chloride, ammonium acetate and aqua each extract, which were then diluted with 5% v/v nitric acid regia. None of the reagents had a significant eVect on the to 20 ml to give solutions containing 1 ng ml-1 of each internal analytical response for uranium, i.e., slopes of the five curves standard. Each solution was then divided into five portions diVered by less than 4%.Direct calibration with respect to which were analysed separately. nitric acid standards was therefore used. The solution detection DiVerent uranium concentrations were obtained with each limit, based on three times the standard deviation of a blank of the internal standards (Table 4) but, since the true concenmeasured ten times divided by the slope of the (5% v/v nitric tration was unknown, it is not possible to comment on the acid) calibration curve, was 0.04 mg l-1.This is equivalent to relative accuracy obtained. Similar precision was obtained for a method detection limit of 8 ng g-1 for extracts and 20 ng g-1 all three internal standards in the analysis of acetic acid for aqua regia digests. extracts. The matrix eVects associated with the other extracts Analysis of sediment samples Table 4 Uranium concentrations measured in sequential extracts using diVerent internal standards (mg g-1 dry weight; n=5) The determination of uranium in sediment digests by ICP-MS and alpha spectrometry is compared in Table 5.Results 115In 209Bi 236U obtained by ICP-MS were generally lower than the corresponding alpha spectrometry values which may have resulted from Mean RSD (%) Mean RSD (%) Mean RSD (%) inaccuracies in the concentration of uranium in the diVerent Step 1 1.68 0.77 1.81 0.61 1.77 0.85 solutions used to calibrate the two techniques.It should also Step 2 0.30 2.33 0.31 1.74 0.31 0.87 be noted that a diVerent digestion method (dry ashing followed Step 3 2.20 1.82 2.10 0.53 2.42 0.46 by open reflux digestion) was used in preparing samples for Residual 1.86 1.94 1.04 10.5 2.25 0.68 the radiometric technique. Nevertheless, the trends in the J. Anal. At. Spectrom., 1999, 14, 163–168 165Table 5 Pseudototal uranium concentrations in sediment digests detection limits. It was not possible to assess rigorously ‘in- (mg g-1 dry weight) batch’ and ‘between-batch’ precision because many of the samples extracted as conventional (in-batch) duplicates were Sample ICP-MSa Alpha spectrometryb found to contain rather low levels of uranium.The results of this study confirm the significant technological CRM 601 1.6±0.1 N/Ac WH1 26±0.2 29±233 enhancement of uranium at Whitehaven previously reported WH2 44±0.3 55±133 by other workers.28–30 Also, a clear decrease in uranium WH3 20±0.2 20±0.633 concentrations has occurred since the Marchon plant ceased WH4 2.2±0.1 3.1±0.133 ore processing in 1992. Uranium concentrations of 20 mg g-1 WH5 9.9±0.1 12±0.433 and above were observed in 1992–93 but levels had decreased WH6 3.5±0.1 3.9±0.133 to an average of around 5 mg g-1 over the subsequent period WH7 6.4±0.1 5.6±0.2 WH8 4.8±0.1 6.6±0.2 from 1994–97.MY1 0.60±0.05 0.74±0.0233 Uranium concentrations in coastal sediments are strongly NT1 0.32±0.05 0.64±0.0233 influenced by local geology, but typical activities in the UK RG1 (0.10)d 0.23±0.0133 are #4–14 Bq kg-1 of 238U,29 equivalent to 0.3–1.1 mg of aICP-MS results are mean values±1 standard deviation, for n=6 U g-1.Hence, despite the decrease, there is still evidence of determinations (i.e., three measurements performed on two samples). technological enhancement in Whitehaven Harbour five years bAlpha spectrometry results represent a single determination±the after the cessation of discharges. Uranium concentrations at counting error.cN/A=not analysed. dThe value in parentheses is Maryport, Nethertown and Ravenglass, however, are within within a factor of ten of the detection limit. the typical ‘background’ range. The distribution of uranium between operationally-defined phases of the Whitehaven sediments was found to vary between two sets of data are similar. A plot of the pseudototal samples (Fig. 1). There also appeared to be correlations concentrations determined by ICP-MS vs.the alpha specbetween speciation, grain-size distribution and concentration. trometry results yields a straight line (equation y=0.82x-0.55; Sediments with uranium concentrations close to typical UK r2=0.99). levels (WH4, WH6, WH7 and WH8) contained a smaller Concentrations of uranium in sediment extracts are given proportion of particles in the <63 mm fraction (see Table 2) in Table 6. Results for CRM 601 are also included, although and a relatively even spread of the analyte in all four oper- the reference material was of no use in validating the present ationally-defined phases. In contrast, residual and, in particu- method because it is not certified for extractable uranium.The lar, oxidisable forms were predominant in the more quality of the sequential extraction was assessed by comparing contaminated sediments (WH1, WH2, WH3 and WH5) which the overall recovery of uranium (Step 1+Step 2+Step 3+ were also characterised by a higher proportion of small Residual ) with the results of aqua regia digestion.Recoveries particles. This would suggest that the decrease in uranium were generally within±10% of the pseudototal concentrations. levels observed over the period of the study is less likely to The high recovery (120%) for RG1, specimen B, may be attributed to the proximity of many of the measurements to have been caused by re-dissolution of uranium and is more Table 6 Uranium concentrations in sediment extracts (mg g-1 dry weight)a Sample Batch Step 1 Step 2 Step 3 Residual Sum Recoveryb (%) CRM 1 Ac (0.02)d <0.008 0.58±0.03 1.13±0.07 1.74 109 B (0.02) <0.008 0.65±0.04 1.22±0.08 1.89 118 2 A (0.02) <0.008 0.50±0.02 1.31±0.05 1.84 115 B (0.02) <0.008 0.53±0.04 1.35±0.08 1.91 119 WH1 1 0.56±0.01 0.24±0.01 14±0.5 10±0.5 25 96 2 0.58±0.01 0.17±0.01 13±0.5 11±0.5 25 96 WH2 1 1.4±0.03 2.6±0.09 33±1.4 11±0.6 48 109 2 1.7±0.04 2.9±0.05 27±1.0 11±0.4 42 95 WH3 1 0.60±0.01 0.22±0.01 14±0.5 5.6±0.3 20 100 2 0.73±0.02 0.21±0.01 13±0.5 5.6±0.2 19 95 WH4 1 0.45±0.01 0.81±0.03 0.70±0.03 (0.14) 2.1 95 2 0.44±0.01 0.84±0.03 0.70±0.03 (0.19) 2.2 100 WH5 1 0.42±0.01 0.32±0.02 8.7±0.5 0.99±0.06 10 101 2 0.45±0.01 0.17±0.02 8.2±0.4 1.4±0.04 10 101 WH6 1 0.43±0.01 0.89±0.03 1.9±0.1 0.23±0.02 3.5 100 2 0.51±0.01 0.98±0.03 1.6±0.1 0.24±0.02 3.3 94 WH7 1 A 0.56±0.01 1.3±0.05 3.3±0.1 0.32±0.02 5.5 86 B 0.54±0.02 1.4±0.05 3.9±0.2 0.36±0.02 6.2 97 2 A 0.65±0.01 1.3±0.03 3.3±0.1 0.45±0.02 5.7 89 B 0.62±0.02 1.2±0.04 3.7±0.2 0.49±0.02 6.0 94 WH8 1 0.64±0.02 1.8±0.10 1.9±0.1 0.41±0.02 4.8 100 2 0.80±0.03 1.6±0.10 1.7±0.1 0.41±0.02 4.5 94 MY1 3 A 0.20±0.01 (0.05) 0.25±0.01 (0.13) 0.63 105 B 0.21±0.01 (0.01) 0.26±0.01 0.20±0.08 0.68 113 WH1 3 A 0.66±0.01 0.24±0.01 17±0.1 5.9±0.01 24 92 B 0.69±0.01 0.26±0.01 13±0.1 8.8±0.05 23 88 NT1 3 A (0.02) (0.02) 0.13±0.01 (0.15) 0.32 100 B (0.03) (0.02) 0.12±0.01 (0.17) 0.34 106 RG1 3 A <0.008 (0.01) (0.03) (0.06) 0.10 100 B (0.02) (0.01) (0.03) (0.06) 0.12 120 aResults are mean values ±1 standard deviation, for n=3.bRecovery=(sum/pseudototal×100); pseudototal values from Table 5. cA and B are replicate specimens extracted in the same batch; Sum=Step 1+Step 2+Step 3+Residual. dValues in parentheses are within a factor of ten of detection limits. 166 J. Anal. At. Spectrom., 1999, 14, 163–168Table 7 Uranium concentrations in ore and phosphogypsum extracts and digests (mg g-1 dry weight)a Sample Step 1 Step 2 Step 3 Residual Sum Pseudototal Alpha spectrometry Ore K 2.03±0.06 2.60±0.04 24.7±0.2 64.0±0.7 93.3 123±1 134±23 Ore Y 2.60±0.04 0.33±0.04 22.5±0.3 68.4±0.5 93.8 112±1 143±12 Phosphogypsum 24.0±0.1 17.7±0.1 5.37±0.06 11.4±0.2 58.5 59±0.5 50±5 aResults are mean values±one standard deviation for n=6 (i.e., three measurements performed on two samples).nation in extracts where concentrations are low. Overall recoveries of uranium by sequential extraction were similar to pseudototal concentrations, but results obtained by ICP-MS appeared to be slightly negatively biased with respect to alpha spectrometry.Application of the method demonstrated the usefulness of sequential extraction for discriminating between potential sources of contamination in the aquatic environment. Further work is in progress to develop methods for the determination of characteristic uranium isotope ratios in extracts and digests,44 which will allow similar studies to be conducted in the vicinity of civil nuclear installations. Acknowledgements S.E.H.acknowledges the EPSRC for provision of a studentship and the organisers of the 9th BNASS meeting for the award of a student bursary. The authors are grateful to Dr A. J. Poole (MAFF, Lowestoft, UK) for providing the samples of phosphate ore and phosphogypsum waste used in the study Fig. 1 Operational speciation of uranium in sediment, ore and and to Katherine Mossop (University of Strathclyde) for phosphogypsum (PPG).analysis of these samples by alpha spectrometry. likely to be a result of dispersion of contaminated sediment References and/or dilution with less contaminated material. 1 M. Kersten and U. Fo� rstner, in Chemical Speciation in the Environment, ed. A. M. Ure and C. M. Davidson, Blackie, Analysis of ore and phosphogypsum samples Glasgow, 1995, pp. 234–271. Specimens of two phosphate ores used by the Whitehaven 2 A.M. Ure, Ph. Quevauviller, H. Muntau and B. Griepink, Intern. J. Environ. Anal. Chem., 1993, 51, 135. plant until 1992, and a sample of phosphogypsum precipitated 3 The certification of the extractable contents (mass fractions) of Cd, from the eZuent stream, were also analysed (Table 7). Cr, Ni, Pb and Zn in sediment following a three-step sequential Pseudototal uranium concentrations in the ores (123 and extraction procedure. CRM 601. Report EUR 17554 EN, 112 mg g-1, equivalent to 1530 and 1390 Bq kg-1, respectively) Commission of the European Communities, Brussels, 1997.were similar to those reported previously by Poole et al.32 The 4 K. Fytianos, S. Bovolenta and H. Muntau, J. Environ. Sci. Health uranium speciation (Fig. 1) was very similar in the two types A, 1995, 30, 1169. 5 Z. Mester, C. Cremisini, E. Ghiara and R. Morabito, Anal. Chim. of ore, with #25% in oxidisable and >70% in residual forms. Acta, 1998, 359, 133.The phosphogypsum (PPG) was characterised by more labile 6 A. U. Belazi, C. M. Davidson, G. E. Keating, D. Littlejohn and species (30% reducible and 41% exchangeable). M. McCartney, J. Anal. At. Spectrom., 1995, 10, 233. It is evident from the results of the study that, where 7 J. Userno, M. Gamero, J. Morillo and I. Gracia, Environ. Int., technological enhancement is most pronounced, the uranium 1998, 24, 487. speciation in the sediment resembles that of the ore more 8 B.Marin, M. Valladon, M. Polve and A. Monaco, Anal. Chim. Acta, 1997, 342, 91. closely than that of the phosphogypsum. This would appear 9 B.Pe� rez-Cid, I. Lavilla and C. Bendicho, Analyst, 1996, 121, 1479. to support the suggestion that spillage of the phosphate ore is 10 M. D. Ho and G. J. Evans, Anal. Commun., 1997, 34, 363. the major contributor to uranium contamination in 11 C. M. Davidson, A. L. Duncan, D. Littlejohn, A. M. Ure and Whitehaven Harbour. Further work is, however, required to L.M. Garden, Anal. Chim. Acta, 1998, 363, 45. assess the potential for redistribution of uranium, in both ore 12 B. R. Harvey, in Chemical Speciation in the Environment, ed. and phosphogypsum, under prolonged exposure to seawater. A. M. Ure and C. M. Davidson, Blackie, Glasgow, 1995, pp. 276–303. 13 V. H. Kennedy, A. L. Sanchez, D. H. Oughton and Conclusions A. P. Rowland, Analyst, 1997, 122, 89R. 14 A. Tessier, P. G. C. Campbell and M. Bisson, Anal.Chem., 1979, A method has been developed for the determination of 51, 844. uranium in sediment digests, and extracts prepared by the 15 K. Bunzl, R. Kretner, P. Schramel, M. Szeles and R. Winkler, BCR, three-stage sequential extraction scheme, by ICP-MS. Geoderma, 1995, 67, 45. 16 M. Trautmannscheimer, P. Schramel, R. Winkler and K. Bunzel, The presence of extractant matrices did not give rise to Environ. Sci. Technol., 1998, 32, 238. interference eVects provided 236U was used as internal standard 17 K.Bunzel, H. Flessa, W. Kracke and W. Schimmack, Environ. and samples were diluted 5-fold with 5% v/v nitric acid prior Sci. Technol., 1995, 29, 2513. to analysis. Detection limits were adequate for quantification 18 A. Baumann, W. Schimmack, H. Steindl and K. Bunzl, Radiat. of uranium at environmental levels in digests and, in most Environ. Biophys., 1996, 35, 229. instances, sediment extracts. However, pre-concentration may 19 K. Bunzl, W. Kracke, W.Schimmack and L. Zelles, J. Environ. Radioact., 1998, 39, 55. be desirable to improve the reliability of uranium determi- J. Anal. At. Spectrom., 1999, 14, 163–168 16720 M. K. Schultz, W. C. Burnett and K. G. W. Inn, J. Environ. 32 A. J. Poole, D. J. Allington, A. J. Baxter and A. K. Young, Sci. Radioact., 1998, 40, 155. Total Environ., 1995, 173/174, 137. 21 I. D. Pulford, R. L. Allan, G. T. Cook and A. B. MacKenzie, 33 G. E. Keating, PhD Thesis, University of Strathclyde, 1996.Environ. Geochem. Health, 1998, 20, 95. 34 M. R. Smith, E. J. Wise and D. W. Koppenaal, J. Radioanal. Nucl. 22 M. S. Baxter, J. Environ. Radioact., 1996, 32, 3. Chem., 1992, 160, 341. 23 G. F. Boothe, Health Phys., 1977, 32, 285. 35 J. Toole, K. McKay and M. S. Baxter, Anal. Chim. Acta, 1991, 24 W. J. Arthur and O. D. Markham, Health Phys., 1984, 46, 793. 245, 83. 25 I. Kobal, S. Brajnik, F. Kaluza and M. Vengust, Health Phys., 36 C. M. Davidson, R. P. Thomas, S. E. McVey, R. Perala, 1989, 58, 81. D. Littlejohn and A. M. Ure, Anal. Chim. Acta, 1994, 291, 277. 26 H. B. Van der Heijde, P. J. Klijn, K. Duursma, D. Eisma, A. J. de 37 Z. Karpas, L. Halicz, J. Roiz, R. Marko, E. Katorza, A. Lorber Groot, P. Hagel, H. W. Ko� ster and J. L. Nooyen, Sci. Total and Z. Goldbart, Health Phys., 1996, 71, 879. Environ., 1990, 90, 203. 38 M. Caddia and B. S. Iversen, J. Anal. At. Spectrom., 1998, 13, 309. 27 A. Martinez-Aguirre, I. Garcia-Orellana and M. Garcia-Leo� n, 39 L. Halicz, M. BarMatthews, A. Ayalon and A. Kaufman, At. J. Environ. Radioact., 1997, 35, 149. Spectrosc., 1997, 18, 175. 28 S. F. N. Rollo, W. C. Camplin, D. J. Allington and A. K. Young, 40 K. Shinotsuka and M. Ebihara, Anal. Chim. Acta, 1997, 338, 237. Radiat. Prot. Dosim., 1992, 45, 203. 41 F. LeCornec and T. Correge, J. Anal. At. Spectrom., 1997, 12, 969. 29 P. McDonald, G. T. Cook and M. S. Baxter, in Radionuclides in 42 J. S. Crain and J. Alvarado, J. Anal. At. Spectrom., 1994, 9, 1223. the Study of Marine Processes, ed. P. J. Kershaw and 43 G. Russ and J. Bazan, Spectrochim. Acta, Part B, 1987, 42, 49. D. S. Woodhead, Elsevier, Amsterdam, 1991, pp. 329–339. 44 S. E. Howe, C. M. Davidson and M. McCartney, Fresenius’ 30 G. E. Keating, M. McCartney and C. M. Davidson, J. Environ. J. Anal. Chem., in the press, November 1998. Radioact., 1996, 32, 53. 31 W. C. Camplin, A. J. Baxter and G. D. Round, Environ. Int., 1996, 22, S259. Paper 8/06778C 168 J. Anal. At. Spectrom., 1999, 14, 163&n
ISSN:0267-9477
DOI:10.1039/a806778c
出版商:RSC
年代:1999
数据来源: RSC
|
10. |
High-performance, flow-based, sample pre-treatment and introduction procedures for analytical atomic spectrometry. Plenary Lecture |
|
Journal of Analytical Atomic Spectrometry,
Volume 14,
Issue 2,
1999,
Page 169-178
Julian F. Tyson,
Preview
|
|
摘要:
High-performance, flow-based, sample pre-treatment and introduction procedures for analytical atomic spectrometry† Plenary Lecture Julian F. Tyson Department of Chemistry, University of Massachusetts, Box 34510 Amherst, MA 01003–4510, USA Received 8th September 1998, Accepted 8th December 1998 Recent and on-going work in the author’s laboratory is described with particular reference to the use of flow injection, continuous flow and HPLC procedures for the development of improved analytical methodology for (a) the determination of trace concentrations of As, Cd, Pb and Se and (b) the determination of various arsenic and selenium compounds.The methods have been applied to the analysis of several diVerent sample matrices, including urine, soil, sediments, waters, plants (garlic, onion, apple leaves), yeast, fruit juices, wine, calcium supplements and marine plankton. The dependence of the LOD of an FI HG method on sample volume is examined and the validity of the proposed rectangular hyperbolic relationship established for a number of diVerent analyses.The use of immobilized tetrahydroborate in conjunction with preconcentration of the analyte on the same anion-exchange resin is described as a possible procedure for improving the LOD. When used in conjunction with ETAAS, an LOD of 0.004 mg l-1 for both As and Se was obtained for a sample volume of 10 ml. The procedure was also used in a method for the determination of inorganic arsensic and methylated arsenic(V) species.Methods for the determination of Pb by HG in the presence of hexacyanoferrate(III) were developed and applied to the analysis of urine, soils, waters and apple leaves. In the case of urine, the interference from the chelating agents used in the treatment of patients with elevated lead was overcome by the addition of Sc. The best LOD of 0.03 mg l-1 was obtained for a procedure in which the lead hydride was trapped on the interior of a flame-heated slotted quartz tube under fuel-lean conditions with subsequent atomization when the flame was made fuel rich (by the injection of a small volume of isobutyl methyl ketone).It has been confirmed that it is possible to determine Cd by a ‘cold vapour’ procedure and it was shown that the nature of the atom cell surface played no part in the atomization process, which appeared to be the spontaneous decomposition of the species evolved from acid solution of cadmium on the addition of sodium tetrahydroborate solution.A modest increase in signal was obtained in the presence of nickel and thiourea. An LOD of 0.02 mg l-1 was obtained for a sample volume of 300 ml and the procedure was used for the analysis of NIST SRM 2711 (Montana Soil ) in which the interference from the high lead content was overcome by coprecipitation with barium sulfate. Three examples of procedures using manifold designs incorporating an ‘eight-port’ rotary valve are given to illustrate the versatility of this component: the separation of high concentrations of uranium (5000 mg l-1) from trace concentrations (1 mg l-1) of Al, Be, Li and Mg for determination by ICP-MS, the automated implementation of the co-immobilization of analyte and tetrahydroborate on an anion-exchange resin and the stopped-flow microwave digestion of human urine for the determination of Se by HG and ETAAS.Improvements in the ion-pair (with trichloroacetate) reversed-phase (C8) HPLC procedure (with ICP-MS detection) for the separation of selenoamino acids (and closely related compounds) were made with a new stationary phase (Waters Symmetryshield RP8) and a small-volume spray chamber. The results of extraction procedures indicated that much of the selenium in yeast and garlic is bound in high molecular mass material.So far only a few of the compounds in the extracts have been identified by retention time matching. A reversed phase (C18), ion-pair (tetrabutylammonium) HPLC procedure for the separation of four arsenic species (arsenite, arsenate, monomethylarsinate and dimethylarsonate), with detection by post-column HG-AAS, has been devised and applied to the extracts of soils spiked with these four species.Low recoveries of arsenite were obtained and microwave energy significantly accelerated the oxidation of arsenite to arsenate. There are many areas of scientific study in which information and cancer-promoting properties of some elements and their about the chemical composition, in terms of the trace or minor relevant compounds.elemental composition of the relevant materials, is needed as In a recent overview of the future of atomic spectrometry part of the problem-solving strategy. A good example of such for environmental analysis, Sturgeon1 identified a number of an area is the study of the biogeochemical cycling of the trends and driving forces which probably have general applicaelements.For some elements, only the most rudimentary bility in the wider context of trace element determinations, knowledge has so far been accumulated, and for others, the namely: (a) more elements will be sought at lower concenrelative magnitude of the contributing processes are unknown. trations, (b) total element determinations are becoming less There are many subsidiary problems to be tackled, such as relevant and speciation is becoming more relevant, (c) there is the elucidation of the chemical basis of the cancer-suppressing a greater need to minimize contamination as a consequence of enhanced instrumental LODs, and (d) there is a need for reduced waste production and reagent consumption.†Presented at the Ninth Biennial National Atomic Spectroscopy Symposium (BNASS), Bath, UK, July 8–10, 1998. Sturgeon also pointed out that ‘Real sample matrices may J. Anal. At. Spectrom., 1999, 14, 169–178 169give rise to a plethora of potential problems, often associated which would be obtained by direct introduction of the sample with high dissolved solids content, including suppression of (as received or as produced by some dissolution and/or digessensitivity, increased background levels and spectral inter- tion procedure).This is particularly so for methods in which ferences. The practical consequence is that, if the eVects cannot the authors wish to use FAAS as the instrumental technique.3 be remedied by simple dilution, matrix separation and concen- This goal is, in itself, reasonable; FAAS is, after all, a well tration of the analyte(s) becomes necessary.This is often the understood technique, is robust and relatively inexpensive, case in any event, as many samples from uncontaminated and requires a minimum of operator skill. By far the most sources contain analytes of interest at concentration levels popular procedure for analyte concentration and matrix which challenge the best instrumental limits of detection’.With separation is solid phase extraction (SPE). In general, the regard to the future, he suggested that ‘Automation of many analyte from a relatively large volume of solution is selectively routine sample preparation tasks can be anticipated. Enhanced retained by a solid reagent phase (by a variety of mechanisms) LODs can be achieved with direct solids (e.g., LA) and slurry and then released into a relatively small volume of eluent. sampling approaches as well as through more extensive use of In addition, chemical vapour generation (CVG), i.e., HG on-line chemical manifolds and FI techniques to minimize and CV procedures, are the procedures of choice4 for the sample contamination and perform matrix separation–analyte determination of As, Se and Hg, and probably also for the concentration functions….Without doubt, the greatest impact determination of Sb, Bi and Ge. In general, the analyte is on sample processing and introduction for atomic spectrometry selectively converted into a volatile derivative which is then has derived from the fields of FI technology and microwave blown out of solution.Both of these procedures give rise to radiation’. an increased signal owing to the greater analyte mass flux Finally, Sturgeon concluded, ‘Current instrumentation is introduced to the atomizer than would be the case for the already capable of providing instrumental detection limits far conventional nebulization of the original sample solution.superior to method detection limits for many elements due to When used for ETAAS (with in-atomizer trapping in the case our inability to control contamination and, hence, the method of CVG), the procedures have the eVect of introducing a blank. It is to be hoped that the widespread use of FI greater mass of analyte into the atomizer than that which techniques and, ultimately, perhaps nanotechnology for sample would have been introduced in a conventional sample volume handling will help alleviate this particular problem’.(say 20 ml ). Again, an increased signal is obtained. It is However, if one takes a broad view of the current research therefore tempting to argue, as many workers do, that this and development eVorts into improving analytical method- increased sensitivity gives rise to increased detection limits and ologies in which commercially available atomic spectrometric that there is a relationship between LOD and sample volume instrumentation is used, then it is apparent that the commonly such that the LOD can be decreased to any desired value; it used techniques have limitations which, in turn, set perform- is simply a matter of preconcentrating a suYciently large ance limits to the methods in which they are incorporated.sample volume. Thus, FAAS does not have LODs low enough, it has a limited working range and limited element coverage and suVers from Detection limits and sample volume matrix interferences; ETAAS suVers from the last three of these limitations, but has instrumental LODs which are so low A recent survey5 (by no means exhaustive) of relevant papers that method LODs are governed by blank contamination; in the area of HG-ETAAS found as many as 15 papers in ICP-OES has a wider elemental coverage (and oVers the which this kind of claim was made.This is not surprising as, possibility of genuine simultaneous multi-element determiin general, it is true that LODs get better as the sample volume nations) and a larger working range but instrumental LODs is increased.However, it is decidedly misleading to give the not good enough for many environmental and clinical appliimpression that there is no limit to how low the LOD can be. cations, and suVers from spectral interferences; and ICP-MS The relationship between these two quantities needs to be has low instrumental LODs (blank contamination becomes a examined a little more closely and, in particular, the situation problem again), with wide elemental coverage and nearly that pertains when an FI procedure is used.simultaneous multi-element determinations but the instrumen- In a recent paper on the determination of Cd by CVG- tation is delicate and suVers from various matrix interference ETAAS with in-atomizer trapping, Geonaga Infante et al. eVects. It is clear from the review literature covering these provided6 the following information: ‘the detection limit was techniques (as exemplified by the Atomic Spectrometry calculated to be 60 ng l-1 for 1.4 ml…of course, lower detec- Updates which appear in the June, August and October issues tion limits could be obtained if higher sample volumes were of this journal ) that much of the published work relating to preconcentrated…7 ml…resulted in a detection limit of the use of these techniques concerns some aspect of the 13 ng l-1’.A scatter plot (see Fig. 1) of these data with LOD limitations listed above. For some techniques, e.g., ETAAS, (on the ordinate) against sample volume, V, (on the abscissa) the majority of technique-based publications are concerned would have two points. The relationship between LOD, CL with achieving a separation between analyte and matrix: either (mg l-1), and V (ml ) could be the straight line, CL=-8.39× by the judicious choice of some thermochemical reactions 10-3V+0.0718. However, this would predict that the LOD within the atomizer, or by the use of powerful background was zero for a sample volume of about 8.6 ml and that for correction procedures, or both.sample volumes greater than this, the LOD would be negative! In essence, the contribution of FI techniques to this area of A more thoughtful analysis of the situation would lead to the analytical instrumental analysis has been to open up a number conclusion that the relationship would be more properly of possibilities for the separation of analyte and matrix with described by one of inverse proportion, i.e., as the signal (for analyte concentration.As these two functions are almost a given concentration) is directly proportional to sample always achieved by appropriate chemical reactions, the use of volume, it might be expected that the LOD would be inversely FI techniques in this fashion has been described2 as ‘putting proportional to V. For the data in Fig. 1, the relationship the chemistry back into analytical chemistry’.would be CL=b/V, where b is a parameter to be fitted. The data, in fact, do not fit this function particularly well as the Flow injection, analyte concentration and detection two values calculated for b are 8.4×10-3 and limits 9.1×10-3 mg ml2. Another iteration of the modelling process is needed. Numerous publications appear each year describing procedures in which the LOD has been improved over that Detection limit is a function of signal-to-noise ratio and 170 J.Anal. At. Spectrom., 1999, 14, 169–178that the major contribution to the a term was the concentration of analyte in the reagents. Results supporting this finding had already been obtained8 for the determination of Se in urine by FI-HG-ETAAS. It was found (a) that the major source of analyte contamination was the borohydride reagent and (b) the best LOD (0.06 mg l-1 in the solution introduced into the manifold) was obtained for an injection volume of 1 ml.However, more data are needed before it may be concluded that for all FI-HG-ETAAS procedures the optimum sample volume is 1 ml. Immobilized tetrahydroborate reagent It may be concluded that a significant improvement in LOD would be obtained for an FI procedure if the reagent volume Fig. 1 Relationship between limit of detection and sample volume for were independent of sample volume. One possibility for HG HG-ETAAS determination of cadmium with in-atomizer trapping.procedures would be to use an immobilized tetrahydroborate Data are taken from ref. 6. The curve is a rectangular hyperbola of reagent. Tesfalidet and Irgum described such a procedure for the form y=b/x+a, where a and b are constants. Details of the the determination of As9 and Se10 in which the tetrahydro- equation are given in the text. borate was immobilized on an anion-exchange resin and the hydride was generated on passage of an acidified sample may be defined7 as the concentration giving a signal equal to solution. It was observed11 that the analyte could also be three times the standard deviation of the blank signal.If the retained on the resin and, as the first stage in developing slope of the calibration is S (in appropriate units) and standard methods based on immobilized tetrahydroborate reagent, the deviation of blank signal is sbl, then the LOD is given by CL= performance of a procedure in which analyte (in this case 3sbl/S.In general, it might be expected that both S and sbl selenite) and the borohydride were co-immobilized was evalu- would be functions of V, with S directly proportional to V for ated. It was found11 that (a) the optimum tetrahydroborate those situations in which the instrument response is directly concentration was 0.05%, considerably lower than values proportional to analyte mass. In general, sbl will be made up typically used in conventional flow or batch procedures, (b) of a contribution from the instrument, s0, and a contribution the procedure gave signals for both AsIII and AsV but only for due to the response to the reagents (including the analyte SeIV (not for SeVI ) and (c) the LOD decreased as the sample contamination in the reagents), sR.volume increased. The procedure was applied to the determi- For a batch procedure, the amounts of reagents are nation of Se in river, lake and tap water; no interferences from independent of sample volume, hence the LOD–V relationship these matrices were encountered for Se concentrations between is CL=3sbl/kV, where k is the constant of proportionality 0.5 and 10 mg l-1. Detection limits were measured for three between sensitivity and sample volume. sample volumes, 3, 6 and 9 ml, for which LODs of 0.24, 0.15 The detection limit can be as low as one wishes, simply by and 0.12 mg l-1, respectively, were obtained.These results are increasing the sample volume, but there is a practical limit set consistent with the FIDL equation CL=0.540/V+0.060 (the by the amount of sample needed and the time taken to process correlation coeYcient of a least-squares fit of a straight line this sample volume.For example, in the determination of Cd to points on a plot of CL versus 1/V is 1.0000). cited above, if this model were valid it would require a sample volume of about 88 ml to obtain an LOD of 0.001 mg l-1; at Analyte concentration by SPE a flow rate of 2.8 ml min-1, it would take about 30 min to process just one sample.As sample processing times become The next stages in the development of this procedure were (a) this long, instrumental drift becomes a contributor to the noise to separate the retention of the tetrahydroborate and the on the signal. analyte, so that the amount of reagent used was independent In a flow-based procedure, whether it involves HG or SPE, of the volume of sample loaded, (b) trap the generated hydride the amount of reagent used is directly related to sample volume on the interior of a graphite furnace and (c) automate the and thus the contribution to sbl from sR increases in direct procedure with a commercially available computer-controlled proportion to the sample volume so that sbl=s0+kV (where flow injection analyzer unit [a Perkin-Elmer (Norwalk, CT, k is the constant of proportionality relating sR and V) and USA) FIAS 200].The separate sample loading step meant CL=3sbl/kV=3(s0+kV)/kV=3s0/kV+3k/k.This equation that the sample volume could be increased substantially withhas the general form CL=b/V+a, and will be referred to as out increasing the blank contribution and it was found12 that the flow injection detection limit (or FIDL) equation. For the for a 20 ml sample volume ( loaded at 20 ml min-1), the LOD data used as example, the parameters a and b may be calculated was 0.004 mg l-1. The amount of tetrahydroborate used per (as two points are available) to give an FIDL equation CL= determination was 5 mg (5 ml of 0.1% m/v solution flowing 8.23×10-3/V+0.00125. This equation is also shown in Fig. 1. at 5.0 ml min-1 for 1.0 min). The LOD was significantly improved over that which would have been obtained for a Detection limits, sample volume and reagent blank sample volume of 20 ml with the quartz tube atomizer and the simultaneous immobilization procedure. Substitution in A more detailed treatment of the problem, which attempted the above FIDL equation gives a value of 0.087 mg ml-1.The to identify the factors contributing to the a and b terms, has procedure was applied13 to the determination of Se in several been presented,5 together with results for the determination of fresh garlic samples. Values ranging from 55 to 350 mg kg-1 As by FI-HG-ETAAS for 10 sample volumes covering the were found. range 0.156–1.56 ml. The data were consistent with an FIDL equation CL=0.041/V+0.007, from which it was concluded, Application to speciation of arsenic compounds with a certain degree of subjectivity based on estimations of the confidence intervals about the LODs, that no significant As the arsenic species AsO(OH)3, As(OH)3, CH3AsO(OH)2 (MMA) and (CH3)2AsO(OH) (DMA) are weak acids with improvement in LOD would be obtained for a sample volumes in excess of 1.00 ml (LOD 0.05 mg l-1).The model indicated suYciently diVerent pKa values, a procedure based on selective J.Anal. At. Spectrom., 1999, 14, 169–178 171after in-atomizer trapping, with a graphite furnace atomizer. The possibility of in-atomizer trapping of lead in a slotted quartz tube in a flame has also been demonstrated. Analyte concentration by SPE Two manifold designs have been used for the concentration of lead as the diethyldithiocarbamate (DDC) complex by retention on a hydrophobic solid phase extractant material. In the first of these,15 the lead DDC complex was formed in the FI manifold by merging the sample stream with a solution of DDC buVered at pH 4–5.The complex was retained on a small column of C18 silica and then eluted with acetonitrile. To avoid the loss of the DDC complex to the walls of the tubing, the confluence point was located inside the loop of the ‘injection’ valve so that in the ‘elute’ position all tubing in contact with the lead DDC complex was washed with eluent. This arrangement is shown in Fig. 3. The manifold was used in the determination of lead in water and fruit juice matrices at concentrations of 0.5 and 5 mg l-1.The enrichment factor, calculated as the ratio of calibration slopes, was about 50 and the detection limit was 6 mg l-1 for a sample volume of 5.8 ml Fig. 2 EVect of sample pH on peak are a signal for four arsenic species and for the blank: A, arsenate; B, arsenite; C, monmethylarsin- (2 min loading). ate; D, dimethylarsonate; and E, blank. The error bars are the The second manifold was simpler.16 The lead–DDC complex standard deviations of five replicate measurements.was formed oV-line at pH 9 by the addition of the minimum amount of a DDC solution to the sample. The lead–DDC complex was retained on the GC stationary phase Chromosorb analyte retention on the anion-exchange resin is possible. The 102 (a styrene–divinylbenzene copolymer) and eluted into a pKa values are AsV, 2.3, AsIII, 9.2, monomethylarsenic(V), 4.0 small volume of ethanol delivered from an injection valve.and dimethylarsenic(V) 6.3, hence at low pH values AsV will This manifold is shown in Fig. 4. An enrichment factor of exist as an anion whereas all the other species will be fully about 25 was obtained with an LOD of 2 mg l-1 for a sample protonated and not retained. The results of the pH study are volume of 4.4 ml (2 min loading). The method was applied to shown in Fig. 2, confirming that this predicted behaviour is the analysis of spiked tap water and artificial sea-water at indeed obtained.A speciation procedure was devised14 based concentrations around 0.1–0.4 mg l-1. A soil SRM (NIST on selective retention of AsV (at pH 2.3) and selective oxidation SRM 2711 Montana Soil, lead content 1162 mg kg-1) was of the other species to AsV. Three stages were needed: in the accurately analysed following digestion with nitric and first, only AsV was determined; in the second, AsIII was hydrofluoric acid in a sealed vessel in a microwave field.selectively oxidized with hydrogen peroxide and nitric acid so No problems either with sample loss to the walls of the the diVerence between the result for this step and that for the tubing or with passage of sample matrix components to the first step gave the AsIII content; in the third stage, all species nebulizer were encountered. This design of manifold would were oxidized to AsV by alkaline peroxodisulfate in a sealed probably not be so suitable for use with ICP instrumentation vessel in a microwave field, so the diVerence between the result which is less tolerant to the delivery of solutions of variable for this step and that for the second step gave the concentration of the methylated arsenic species.A fully automated system was constructed around a Perkin- Elmer FIAS 200 unit interfaced to an ETA atomic absorption spectrometer and the method applied to the determination of arsenic species in river and tap water.The sample was loaded at 10 ml min-1 for 60 s, followed by 0.2% tetrahydroborate at 5 mlmin-1 for 60 s. The hydride was generated on the passage of 8 M HCl at 5 ml min-1 for 45 s. The LOD for a 10 ml sample volume was 0.004 mg l-1. This value may be compared (a) with the value of 0.04 mg l-1 obtained for previous FI-HGETAAS determination of As for which a sample volume of 0.500 ml was used5 and (b) with the predicted value from the relevant FIDL equation for a 10 ml sample of 0.011 mg l-1, showing that a significant decease in the contribution of the blank had been obtained by using the anion-exchange resin.A reference water sample (total As content 80 mg l-1) was accurately analysed and the interferences of nine cations and eight anions on 0.500 mg l-1 As were investigated. FI determination of lead Fig. 3 Solid phase extraction manifold with confluence point in sample loop of six-port rotary valve. In the ‘load’ position (a), the sample, S, Two approaches to the determination of lead have been under is merged with buVered DDC solution and the resulting complex is investigation.In the first, lead has been concentrated on a retained on the column, COL. In the ‘inject’ position (b), pump P1 is solid phase extractant and then eluted into a smaller volume oV while pump P2 keeps running, delivering eluent, E. The lead–DDC of eluent for determination by FAAS. In the second approach, complex is back-flushed from the column to the spectrometer, FAAS, lead hydride has been generated, separated from solution and and any complex adhering to the walls of the tubing is also dissolved.W is waste. determined by AAS either with a quartz tube atomizer or, 172 J. Anal. At. Spectrom., 1999, 14, 169–178findings are in agreement with those of several other workers.20–23 A method was developed for the determination of lead in human urine.24 After optimization of the various flow injection parameters, the LOD was 0.08 mg l-1 for a sample volume of 0.500 ml.A quartz tube atomizer, a Nafion dryer (between the gas–liquid separator and the atomizer) and peak height quantification were used. An RM [NIST SRM 2670 Trace Metals in Urine with Pb concentrations of 10 mg-1 (normal ) (a suggested value) and 109 mg l-1 (elevated) (a certified value)] was accurately analysed.25 However, when the possible interference of a chelating agent, such as EDTA (used in the treatment of patients with elevated lead), was investigated it was clear that the simple sample prepreatment of addition of hexacyanoferrate(III) and dilution with hydrochloric acid solution would not be suitable for the analysis of ‘real’ urine samples.The method was modified by the addition of Sc, whose EDTA complex formation constant is 105 times larger than that for PbEDTA, and the lead signal was restored. The modified procedure was applied blind to the analysis of 50 samples provided by the New York State Department of Health.Samples were diluted between 5- and 100-fold, and the solution injected contained 2.5×10-4 M Sc, 5% m/v hexacyanoferrate(III) and 0.1% HCl. Calibration with aqueous Fig. 4 Solid phase extraction manifold with Perkin-Elmer four-port standards was possible. A comparison of the results of the rotary valve. In the ‘load’ position (a), the buVered sample and DDC determinations by FI-HG-AAS with those obtained by the are delivered to the column, COL.In the ‘inject’ position (b), pump New York State Department of Health (in whose laboratory P1 is oV while pump P2 keeps running, delivering a discrete volume an ETAAS procedure was used) is shown in Fig. 5, from of eluent, E, via carrier C. The lead–DDC complex is flushed through which it can be seen that there is no significant diVerence at the column to the spectrometer, FAAS. W is waste. the 95% confidence level between these two sets of results. The procedure has been further developed by trapping the acidity and high dissolved solids.In these cases, a manifold hydride at 300 °C on the interior of a graphite furnace atom- design in which the matrix components were diverted to waste izer.25 The atomizer was first pretreated with Ir (120 mg). The during loading would be more suitable. LOD was 0.12 mg l-1 for a sample volume of 1.00 ml. However, after purification of the hexacyanoferrate(III) by passing a Hydride generation solution through a cation-exchange resin, the LOD was decreased to 0.03 mg l-1.Unfortunately, this procedure did There has been relatively little interest in the HG determination of Pb in comparison with that in the HG determination of As not prove viable for routine use as the capacity of the resin to retain lead rapidly became exhausted. The procedure was used and Se, for which HG-AAS is probably now the procedure of choice. Unlike the situation for the hydride-forming elements in a method for the determination of Pb in calcium supplements.Two materials (CVS brand) were examined: one of Groups 14 and 15 of the Periodic Table, lead hydride is formed from a precursor in the highest oxidation state in contained 333 mg Ca+133 mg Mg+5 mg Zn per tablet and the other contained 500 mg Ca (from crushed oyster shells) solution, not the lowest. Even with an oxidative sample digestion or dissolution, lead is still in the +2 state and per tablet. In terms of the lead content, the former contained 0.57 mg kg-1 (i.e., 0.43 mg per tablet) and the latter eYcient generation of plumbane is only achieved in the presence of an oxidant.There are, therefore, a large number of 0.66 mg kg-1 (i.e., 0.43 mg per tablet). The method was valipossible reactions. Lead may participate in an oxidation reaction (from PbII to PbIV by the added oxidant), a hydride transfer reaction (the formation of plumbane from PbIV ), and reduction reactions (tetrahydroborate is a strong reductant and also a hydride transfer reagent and can reduce PbIV to PbII and even to Pb0, which in turn may be oxidized by protons from the acid).In addition to the primary hydride generation reaction, tetrahydroborate reacts with acid (to form boric acid and hydrogen) and may react with the added oxidant in a redox reaction (most likely forming boric acid and hydrogen). It is perhaps surprising that any plumbane is generated at all, given the delicate balancing needed between the thermodynamics and kinetics of all of these possible reactions.In addition, some workers have found it advantageous to add a complexing agent (such as malate, tartrate or lactate), notionally to stabilize the PbIV species. The conflicting literature regarding Pb HG has recently been summarized.17,18 Various conditions for the generation of plumbane have Fig. 5 Scatter plot of data for the determination of lead in urine by been investigated,15,19 resulting in the conclusion that the (a) HG-AAS (ordinate) and ETAAS (abscissa).The line is the greatest sensitivity for FI-HG-AAS with quartz tube or graph- unweighted least-squares linear regression fit to the data and has slope ite furnace atomization is obtained in the presence of hexa- 0.963±0.041, intercept 9.17±18.4 and correlation coeYcient 0.992. The ±terms are 95% confidence intervals. cyanoferrate(III) as ‘oxidant’ and hydrochloric acid. These J. Anal. At. Spectrom., 1999, 14, 169–178 173dated by the analysis of the materials by ETAAS methods Chemical vapour generation (three diVerent sample pre-treatment procedures were used) When tetrahydroborate is introduced into an acidified solution and by spike recoveries. Further studies of the HG of lead are of cadmium, a volatile derivative is formed which can be either in progress as the LOD for ETAAS is no better than that for transported into a quartz tube atomizer in the light path of quartz tube atomization, although the sensitivity of the former an atomic absorption spectrometer or trapped on the interior is much higher.It is concluded that the noise associated with of a graphite furnace atomizer. The absorption signal produced the ETAAS procedure is much higher, and as the only signifi- from either atomizer is proportional to the cadmium concen- cant diVerence between the two procedures is the in-atomizer tration in the original solution and this phenomenon forms trapping stage, it is thought that this process is, for some the basis of a viable method of analysis for cadmium.Sanz- reason as yet unknown, less reproducible for plumbane than Medel and co-workers33 made the interesting observation that, for other hydrides such as arsine and hydrogen selenide. with the quartz tube atomizer, the procedure worked satisfac- Clearly improved LODs would be obtained with purer hexatorily even when the tube was at room temperature. They cyanoferrate(III) and work is in progress to find a routine provided convincing spectroscopic evidence34 that the absorb- method for the removal of the lead from this reagent.ing species were indeed cadmium atoms and they proposed, Finally, in this saga of method development based on therefore, that the procedure be called ‘cold vapour’ by analogy plumbane generation, a procedure based on trapping the with the corresponding procedure for the determination of hydride on the interior of a slotted quartz tube atomizer has Hg.It has been postulated that the species evolved from been devised.26 The concept of trapping atom precursors on a solution is cadmium hydride (CdH2), but that this is suYciently cool surface in a flame was first described by Lau et al.27 and unstable that a significant proportion has decomposed to the concept of increasing atom residence time in the flame by cadmium atoms during the time taken to pass from the the insertion of a slotted quartz tube was first described by gas–liquid separator to the atomizer.As cadmium atoms are Watling.28 The combination of the two procedures was pronot thermodynamically stable at room temperature with posed by Ataman and co-workers,29,30 who showed that it respect to the bulk metal, the cadmium atoms formed in this was possible to trap the precursor species in a fuel-lean flame fashion are presumably kinetically ‘stable’. These results have and atomize in a fuel-rich flame produced either by rapidly changing the fuel flow or by the pulse nebulization of an been confirmed by Guo and Guo35 and Kradtap.15 organic solvent.They also showed that it was possible to The best conditions for the generation of the precursor of dispense with the water-cooled tube and trap on the slotted the atoms are the subject of some debate. It has been reported tube.31 This concept has been extended26 by the introduction that (a) a surfactant should be present,34 (b) that a considerable of the lead as the hydride into the slotted quartz tube in a signal enhancement occurs in the presence of nickel, cobalt fuel-lean flame. Essentially the same generation chemistry was and thiourea35 and (c) that potassium tetrahydroborate gives used as for the previously developed HG methods. The lead a significantly enhanced signal compared with that obtained was atomized on the injection of 50 ml of isobutyl methyl in the presence of sodium tetrahydroborate.36 All of these ketone.The LOD was measured for three diVerent sample claims have been investigated in some detail37 for a typical volumes, 2.6, 5.2, and 7.8 ml, for which the LODs were 0.075, two-line flow injection system (i.e., the acidified sample is 0.047 and 0.028 mg l-1, respectively. These data fit an FIDL injected into an acid carrier which merges with an alkaline plot of the equation CL=0.175/V+0.009 with a correlation tetrahydroborate stream; after passage through a reaction coil, coeYcient of 0.99.The procedure was used as part of a method argon is merged and the mixture flows through a stripping for the determination of Pb in two SRMs: NIST SRM 2709 coil to the gas–liquid separator). The findings are that a San Joaquin Soil (19 mg kg-1) and SRM 1515 Apple Leaves surfactant has no eVect, nor has potassium tetrahydroborate. (0.47 mg kg-1) were analysed accurately. In the presence of 1% thiourea and 10 mg l-1 Ni or Co, the Work is in progress32 on the determination of lead by a peak height signal is enhanced by 10–20% but a black precipicombination of SPE (the DDC complex on Chromosorb 102) tate rapidly forms in the presence of Co.In addition, it has with HG from the column eluent. The PbDDC was eluted been found that the surface of the atom cell plays no role in with 0.3% ethanolic hydrochloric acid solution and merged the atomization process. A Pyrex glass cell of smaller diameter with a hexacyanoferrate(III ) stream prior to merging with than the ‘standard’ quartz cell can be used to increase the tetrahydroborate. For quartz tube atomization, the LOD was peak height signal.However, when a Nafion dryer was used 0.2 mg l-1 (i.e., about a factor of 10 improvement over the to transport the volatile species to the atom cell, the signal value obtained for elution with ethanol and determination by was lost completely. A small open chamber, of volume about FAAS). The procedure was used as part of a method for the 5 ml, was used as a gas–liquid separator, which diverted only determination of Pb in wine over the concentration range 5 8% of the cadmium to waste.It was observed that better -25 mg l-1. These results are interesting from two points of signals were obtained if the acid concentration in the injected view: first, it appears that plumbane can be generated in the sample was lower than that in the carrier and that the length presence of DDC (the acid concentration may be high enough of tubing between the gas–liquid separator and the atom cell to release the lead from the complex) and ethanol, and second, should be the minimum value possible.All of the evidence it appears that DDC is capable of removing lead from the suggested that the signal characteristics were due to the kinetic various complexing agents in the wine. features of an unstable analyte species. Interestingly, peak height was a more precise and robust quantitative parameter than peak area.Determination of cadmium Procedures were developed for the accurate determination of Cd in NIST SRM 2711 Montana Soil at 41.7 mg kg-1 and The SPE chemistry, based on the retention of the DDC NIST SRM 1515 Apple Leaves at 0.013 mg kg-1. The high complex on C18 with elution by acetonitrile, has also been lead content of Montana soil (1162 mg kg-1) interfered and used for the determination of Cd in waters and fruit juices.15 was removed by coprecipitation with barium sulfate.In pres- Detection limits were measured for three diVerent sample ence of thiourea and Ni, the slopes of the standard additions volumes, 0.79, 2.36 and 7.08 ml, for which values of 1.3, 0.65 calibration and that for aqueous standards were not signifi- and 0.39 mg l-1, respectively, were obtained. These data fit cantly diVerent. The LOD was 0.016 mg l-1 (only one value the FIDL equation CL=0.799/V+0.289 with a correlation coeYcient of 0.999.for a sample volume of 0.300 ml has been obtained so far). 174 J. Anal. At. Spectrom., 1999, 14, 169–178The eight-port flow injection valve While a variety of FI apparatus has been used in the experiments described above, the automation of several procedures has been greatly facilitated by the use of a twoposition ‘eight-port’ rotary valve. This designation is somewhat of a misnomer, as the device actually has 16 ports, eight on the stator and eight on the rotor.To illustrate the versatility of the device, three diVerent uses are shown. First, the device has been used in a manifold for determinations of Se and As by HG-ETAAS in which the hydrides have been generated on the passage of acid through an column of anion-exchange resin on which the analyte and tetrahydroborate have been sequentially immobilized. The manifold and its operation are explained in Fig. 6. Second, a manifold for the separation of analyte from matrix by retention of the matrix on a solid phase extractant has been designed around this device.The manifold has been used for the separation of uranium (at concentrations up to 5000 mg l-1) from solutions of light elements (Al, Be, Li and Mg), allowing their determination at concentrations down to 1 mg l-1 by ICP-MS.38 The manifold and its operation are shown in Fig. 3 in ref. 38. Third, the device has been used in a manifold for the on-line microwaveassisted digestion of selenium compounds in urine, allowing the determination of Se in this matrix by HG.It has been shown8 that it is possible to convert all Se compounds in human urine to selenite, by heating under reflux in the presence of bromine generated from the reaction of bromate with hydrobromic acid, but attempts to adapt this sample pretreatment to an on-line format had proved unsuccessful in terms of producing conditions which would convert trimethyl- Fig. 7 Manifold for stopped-flow, closed-loop, microwave digestion.With the valve and pumps as shown in (a) the injection loop is loaded with sample solution. The pumps and valve are switched to the positions shown in (b) and the sample solution delivered by an acid carrier, A, into the digestion coil in the microwave cavity. When the valve is switched to the position shown in (c) and both pumps are oV, the sample is digested in a sealed closed loop. The valve and pumps are then activated as shown in (b) and the digested sample pumped via the cooling coil to the HG part of the manifold and merged with a tetrahydroborate stream, B.Intermediate stages in which the probe is introduced into and removed from the graphite furnace are not shown. selenonium (TMSe) (the Se species found in human urine most resistant to oxidative attack) into selenite.25 However, recently it has been found39 that a manifold design in which the sample zone is isolated in a closed loop in the microwave field produces conditions under which TMSe is converted into SeIV. The manifold design is shown in Fig. 7. Under the same operating conditions, SeVI, selenocystine, selenomethionine and selenoethionine were also quantitatively converted into SeIV. Speciation of selenium and arsenic There is considerable research activity directed towards the development of methods for the separation and quantification of the various inorganic and organaometallic species of arsenic Fig. 6 Manifold for HG after preconcentration and immobilization and selenium found in environmental and clinical samples.of tetrahydroborate. With the pumps and valve as shown in (a), Numerous chromatographic (mainly HPLC) and nonsample, S, is loaded on to the column. The pumps and valves are then chromaotgraphic procedures have been published, many of switched to the positions shown in (b), allowing tetrahydroborate, B, which incorporate AAS or ICP-MS detection.40,41 However, to be loaded on to the column. In the final stage, the pumps are it appears that (a) is it diYcult to achieve eYcient (>1000 switched as shown in (c) and the hydride generated on passage of the plates) LC separation of inorganic and methylated arsenic acid stream, A.Various intermediate stages, in which wash solution, C, removes residual material, are not shown. W is waste. species in short times (<5 min) and (b) many clinical and J. Anal. At. Spectrom., 1999, 14, 169–178 175biological materials contain numerous organoselenium com- procedure was validated by the analysis of various reference materials.49 Although these studies are far from complete, pounds which are poorly resolved at short elution times or are unidentified, or both. some general observations have been made: for most materials studied a greater proportion of the total Se was removed by the enzymatic extraction than by the other procedures, suggest- Selenium speciation ing that most selenium is incorporated into larger molecular As part of a broad-based study of the biogeochemical structures that are broken down by the enzyme.The species transformations of selenium compounds and of the nature of profile depends on the material (not all yeasts give the same selenium compounds in materials with anti-cancer properties, profile, for example) and the profile also depends on the analytical methods are under development in which selenium concentration of the selenium. At this stage it diYcult to draw compounds have been separated by HPLC and detected by any conclusions about the relative amounts of various selenium ICP-MS.As the bulk of the studies so far have been concerned compounds which occur at natural concentrations. It has also with plant and yeast materials, the targets have been amino been observed that for several diVerent materials (including acids and closely related species. Recent eVorts have been yeast, garlic, plankton, Brazil nuts and bacteria), the maximum devoted to improving the interface between the chromatograph amount of selenium that the organism can tolerate is about and the spectrometer and to improving the performance of 2 000 mg kg-1.the chromatographic separation. Initial studies were made of A number of potential problems for the analytical ion-exchange separations and reversed-phase separations of methodology have also been identified: recovery of the selamino acid derivatives.42,43 However, an ion-pair reversed- enium species from the column is a function of concentration, phase procedure44,45 had superior performance in terms of (a) ranging from about 60% at 0.1 mg l-1 to over 90% at chromatographic separation, (b) ease of interfacing with the 100 mg l-1.Sensitivity is a function of species, with a 28% mass spectrometer and (c) the application of this procedure diVerence between the most sensitive and the least sensitive to the analysis of a variety of sample materials.compounds and, in the determination of total Se, the sensitivity Further modifications of the conditions were made46 so that is a function of the acid concentration remaining after the the composition of the mobile phase was 1% MeOH+0.6% digestion is complete. This problem has been noted by other trifluoroacetic acid (TFA)+25 mg l-1 Ge (as internal stan- workers and it may be overcome to a large extent by the dard) adjusted to pH 2 with ammonia solution and the station- addition of a mixture of water-soluble tertiary amines (known ary phase was Waters SymmetryShield RP8.The improvement as CFA-C amines).50 in chromatographic performance is shown in Fig. 8 for four Initial work on the identification of the selenium compounds selenium standard compounds. The injection volume was 10 ml has been based on retention time matching with standards.43–45 and the mobile phase flow rate was 1.0 ml min-1. Also con- However, most of the compounds are still to be identified as tributing to the improved resolution were the use of a no match with existing standards has been found.More Meinhard nebulizer and a spray chamber of volume approxi- recently, the approach of derivatizing with reagents designed mately 14 ml containing a fixed 1 cm diameter glass impact to identify particular chemical functionalities has been bead. attempted. These reagents have included hydrogen peroxide Three diVerent extraction procedures have been used.The (as an oxidant), thiosulfate (as a reductant) and methanol or first of these was a simple hot water extract as it is known47 ethanol with sulfuric acid (as an esterifying agent). In this that in the case of yeasts and allium vegetables, anti-cancer procedure, the chromatograms before and after reagent action is obtained with such extracts. The second was a water addition are compared to identify (a) which compounds may extract in the presence of an enzyme (protease XIV)48 which be oxidized or reduced (and whether the process is reversible) reduces polypeptides to dipeptides and the third extraction or (b) which compounds may be esterified.Future studies51 was with hydrochloric acid (0.1 M)+ethanol (95%). In each will involve the use of LC-MS (electrospray ionization with case the extraction was carried out for several hours, and the an ion trap spectrometer) for which initial studies on known extract was filtered through a 0.45 mm filter and a 10 000 selenium compounds indicate that the spectral interpretation molecular mass cut-oV filter (to remove polymeric material, may not be so complicated (as the presence of Se in a fragincluding excess enzyme reagent).Total Se in the original ment ion is easily deduced from the characteristic isotope material and in the residues after extraction and filtration was pattern), but that the sensitivity of this form of LC-MS is determined by a procedure in which the material was digested comparatively low.with nitric acid in a sealed vessel in a microwave field. This Arsenic speciation A common preservative for structural timber is chromated copper arsenate. The timber is treated under pressure with an aqueous solution of the oxides. As part of an on-going study into the fate of arsenic leached from pressure-treated timber into soils, a method for the determination of AsIII, AsV and the two methylated forms of arsenate (discussed earlier) in soils has been under development.The hypothesis to be examined is that the inorganic arsenic which gets into the soil is transformed to volatile methylated species by the action of soil micro-organisms. Initially, an anion-exchange HPLC separation was devised52 and it was shown, on the basis of ultrasound-assisted extraction with methanol+hydrochloric Fig. 8 Comparison between original and modified conditions for the acid (50%+10%), that an arsenic-containing reference mateseparation of selenium compounds.The original conditions41 were a rial, NIST SRM 2704 BuValo River Sediment, contained Zorbax SB-C8 column with 2% methanol+0.1% trifluoroacetic acid arsenite and dimethyl arsinate. Following the success of the as mobile phase at a flow rate of 1.0 ml min-1. The modified conditions ion-pair procedure for the separation of selenium amino acids, (see text) were a Symmetryshield RP8 column with 1% methanol+0.6% work has been directed53 to the development of a similar trifluoroacetic acid (adjusted to pH 2).Peaks: 1, selenate; 2, selenite; 3, selenocystine; and 4, Se-methyl-DL-selenocysteine. procedure for the separation of these four arsenic species. A 176 J. Anal. At. Spectrom., 1999, 14, 169–178performance characteristic and the relevant information should be included in publications. There are distinct possibilities for the use of HG in methods for the determination of Pb and Cd, and such procedures could become as popular as those for the determination of As and Se.The generation of lead hydride is greatly facilitated by the presence of hexacyanoferrate(III). Cadmium may be determined by AAS at room temperature. For the determination of various selenium and arsenic species in environmental and food samples, separation by reversed-phase ion-pair HPLC with element-specific detection seems to be a viable stage in the overall method, despite some of the diYculties encountered related to recovery from the column and the dependence of sensitivity on chemical factors.Fig. 9 Elution profile for arsenic species separated by reversed phase ion-pair HPLC with HG-AAS detection. The elution order is arsenite, In the case of the selenium compounds, the identification of dimethylarsonate, monomethylarsinate and arsenate. The flow rate species responsible for large numbers of peaks in the chromatowas 1.0 ml min-1. grams remains a major area of further work. For both As and Se compounds, the problems of extraction of the compounds from solid matrices also remain as a major area for further work.procedure had been devised based on the use of a Waters Resolve C18 column with 6 mM potassium dihydrogen- Many helpful discussions with the following collaborators phosphate+2 mM tetrabutylammonium hydroxide+0.2% and co-workers are gratefully acknowledged: Peter Uden MeOH at pH 5.7 (adjusted with phosphoric acid) as the mobile (University of Massachusetts, Amherst, MA), Eric Block phase.For detection by post-column HG-AAS with quartz (SUNY, Albany, NY), Helen Crews (MAFF, UK), Erik tube atomization, LODs between 0.5 and 2 mg l-1 have been Larsen (Danish Veterinary and Food Administration), Latif obtained for a 100 ml injection volume. The chromatographic Elci (Erciyes University, Turkey), Nusret Ertas, (Middle East profile is shown in Fig. 9. Technical University, Ankara, Turkey), Patrick Parsons, (New The extraction of these arsenic species from soil is currently York State Department of Health), Walter Slavin (Bonaire being studied.53 Obtaining quantitative recovery without trans- Technologies), Pablo Carrero (University of the Andes, formation is proving to be diYcult.A variety of extractants Merida, Venezuela) and Susan McIntosh, Chris Hanna, Frank have been used, including tetrabutylammonium hydroxide and Fernandez, Glen Carnrick, Gerhard Schlemmer and Eric phosphoric acid (at a variety of pH values with potassium Denoyer (Perkin-Elmer). In addition, the eVorts of the follow- dihydrogenphosphate as buVering agent). The possible roles ing graduate students, who conducted much of the experimen- of ultrasound and microwave energies are also being studied.tal work described above, is gratefully acknowledged: Zikri So far it may be concluded that it is possible to extract MMA Arslan, Susan Bird, Robert Ellis, Hakan Gurleyuk, Mihaly and DMA, but AsIII is strongly bound immediately on spiking Kotrebai, Supaporn Kradtap, Nils Sundin, Cesar Vargas and into a soil matrix and, so far, it has not proved possible to Peter Yehl.recover all of this species. The application of microwave energy results in the conversion of some AsIII into AsV and therefore some of the earlier good recoveries for AsV are called into References question. This is a diYcult analytical problem and there is still 1 R. E. Sturgeon, J. Anal. At. Spectrom., 1998, 13, 351.a long way to go with this method development. 2 J. F. Tyson, Microchem. J., 1992, 45, 143. 3 S. J. Hill, J. B. Dawson,W. J. Price, I. L. Shuttler and J. F. Tyson, J. Anal. At. Spectrom., 1997, 12, 327R. Conclusions 4 J. Dedina and D. Tsalev, Hydride Generation Atomic Absorption Spectrometry,Wiley, Chichester, 1995. Flow-based procedures, in which sample pre-treatment 5 J. F. Tyson, R. I. Ellis, S. A. McIntosh and C. P. Hanna, J. Anal. chemistry is performed in narrow-bore conduits, have a At.Spectrom., 1998, 13, 17. number of characteristics which can lead to the development 6 H. Goenaga Infante, M. L. Fernandez-Sanchez and A. Sanzof improved analytical methods. In addition to providing a Medel, J. Anal. At. Spectrom., 1996, 11, 571. closed, contamination-free environment for the handling of 7 Analytical Methods Committee, Analyst, 1987, 112, 199. 8 J. F. Tyson, N. G. Sundin, C. P. Hanna and S. McIntosh, samples, flow procedures allow for the automated or semi- Spectrochim.Acta, Part B, 1997, 52, 1773. automated implementation of analyte and matrix separations 9 S. Tesfalidet and K. Irgum, Anal. Chem., 1989, 61, 2079. that would be tedious, time consuming and prone to contami- 10 S. Tesfalidet and K. Irgum, Fresenius’ J. Anal. Chem., 1991, 341, nation if performed in a batch mode. In addition, many flow 532. procedures allow for simple direct coupling between the sample 11 P. E. Carrero and J. F. Tyson, Analyst, 1997, 122, 915.pre-treatment stages and the instrumental measurement stage 12 P. E. Carrero and J. F. Tyson, Spectrochim. Acta, Part B, in the press. of the overall method. Flow procedures are a good way to 13 P. E. Carrero and J. F. Tyson, Spectrochim. Acta, Part B, submit- implement solid phase extraction and chemical vapour generted for publication. ation procedures, although the usual increase in reagent con- 14 P. E. Carrero and J. F. Tyson, J. Anal. At. Spectrom., submitted sumption as the sample volume increases produces a for publication.rectangular hyperbolic relationship (the so-called FIDL equa- 15 S. Kradtap, MS Thesis, University of Massachusetts, Amherst, tion) between sample volume and LOD which is not asymptotic MA, 1996. 16 L. Elci and J. F. Tyson, work in progress. to the concentration axis. There are, therefore, limits to the 17 J. Dedina and D. L. Tsalev, Hydride Generation Atomic Absorption extent to which LODs can be improved by increasing the Spectrometry,Wiley, Chichester, 1995, p. 288. sample volume. However, in the case of chemical vapour 18 C. Camara and Y. Madrid, Analyst, 1994, 119, 1647. generation, the use of an immobilized reagent removed the 19 R. E. Ellis, L. Elci, N. Ertas, C. Vargas and J. F. Tyson, work dependence of reagent blank on sample volume and a way in progress. has been opened up for further improvements in detection 20 R. Thao and H. Zhou, Fenxi Huaxue, 1985, 13, 283. 21 P. Zhang and Z. Hu, Fenxi Huaxue, 1987, 15, 404. limits. The dependence of LOD on sample volume is a useful J. Anal. At. Spectrom., 1999, 14, 169–178 17722 H. Chen, J.Wu and I. D. Brindle, Talanta, 1994, 42, 353. 38 P. E. Becotte-Haigh, J. F. Tyson and E. R. Denoyer, J. Anal. At. 23 H. O. Huag, Spectrochim. Acta, Part B, 1996, 51, 1425. Spectrom., 1998, 13, 1327. 24 J. F. Tyson, R. I. Ellis, N. Ertas, F. Fernandez and 39 P. E. Carrero and J. F. Tyson, work in progress. S. A. McIntosh, Clin. Chem., submitted for publication. 40 S. J. Hill, J. B. Dawson, W. J. Price, I. L. Shuttler, C. M. M. Smith 25 R. I. Ellis, PhD Dissertation, University of Massachusetts, and J. F. Tyson, J. Anal. At. Spectrom., 1998, 13, 131R. Amherst, MA, 1997. 41 J. R. Bacon, J. S. Crain, L. Van Vaeck and J. G. Williams, J. Anal. 26 N. Ertas, Z. Arslan and J. F. Tyson, work in progress. At. Spectrom., 1998, 13, 171R. 27 C. M. Lau, A. Held and R. Stephens, Can. J. Spectrosc., 1976, 42 H. Ge, PhD Dissertation, University of Massachusetts, Amherst, 21, 100. MA, 1997. 28 R. J.Watling, Anal. Chim. Acta, 1977, 94, 181. 43 S. M. Bird, H. Ge, P. C. Uden, J. F. Tyson, E. Block and 29 O. Y. Ataman, N. Ertas, R. S. Helles and S. Kumser, Paper E. Denoyer, J. Chromatogr. A, 1997, 789, 349. presented at the FACSS XX Annual Conference, Detroit, MI, 44 S. M. Bird, P. C. Uden, J. F. Tyson, E. Block and E. Denoyer, October 17–22, 1993, Paper 267. J. Anal. At. Spectrom., 1997, 12, 785. 30 N. Ertas, S. Kumser, D. Karadeniz, R. S. Helles and 45 S. M. Bird, M. Kotrebai, P. C. Uden, J. F. Tyson, E. Block and O. Y. Ataman, Paper presented at the FACSS XXIII Annual E. Denoyer, Fresenius’ J. Anal. Chem., 1998, 362, 447. Conference, Kansas City, MO, September 29–October 4, 1996, 46 M. Kotrebai, P. C. Uden and J. F. Tyson, work in progress. Paper 607. 47 C. Ip and D. J. Lisk, Carcinogenesis, 1995, 16, 2649. 31 N. Ertas, D. Karadeniz and O. Y. Ataman, Paper presented at the 48 N. Gilon, M. Potin-Gauthier and M. Astruc, J. Chromatogr. A, XXX Colloquium Spectroscopicum Internationale, Melbourne, 1996, 750, 327. September 21–26, 1997, Paper C66. 49 S. M. Bird, PhD Dissertation, University of Massachusetts, 32 L. Elci and J. F. Tyson, work in progress. Amherst, MA, 1998. 33 A. Sanz-Medel, M. R. Fernandez de la Campa, M. C. Valdez- 50 A. Krushevska, M. Kotrebai, A. Lasztity, R. M. Barnes and Hevia and M. Bordel, Anal. Proc., 1995, 32, 49. D. Amarsiriwardena, Fresenius’ J. Anal. Chem., 1996, 355, 793. 34 A. Sanz-Medel, M. R. Fernandez de la Campa, M. C. Valdez- 51 I. Kaltashov, M. Kotrebai, P. C. Uden and J. F. Tyson, work Hevia and M. Bordel, Anal. Chem., 1995, 67, 2216. in progress. 35 X.-W. Guo and X.-M. Guo, Anal. Chim. Acta, 1995, 310, 377. 52 P. M. Yehl and J. F. Tyson, Anal. Commun., 1997, 34, 49. 36 H. Matusiewicz, M. Kopras and R. E. Sturgeon, Analyst, 1997, 53 H. Gurleyuk, P. C. Uden and J. F. Tyson, work in progress. 122, 331. 37 C. Vargas-Razo and J. F. Tyson, J. Anal. At. Spectrom., submitted for publication. Paper 8/07025C 178 J. Anal. At. Spectrom., 1999, 14, 169–178
ISSN:0267-9477
DOI:10.1039/a807025c
出版商:RSC
年代:1999
数据来源: RSC
|
|