Fresenius J Anal Chem (2001) 370 : 553–558
© Springer-Verlag 2001
TECHNICAL NOTE
M. M. Castiñeira · R. Brandt · A. von Bohlen · N. Jakubowski
Development of a procedure for the multi-element determination of trace elements in wine by ICP–MS
Received: 9 February 2001 / Revised: 23 March 2001 / Accepted: 26 March 2001
Abstract An inductively coupled plasma mass spectrometric (ICP–MS) procedure has been developed for the determination of trace elements in wine. The procedure consists in simple 1+1 dilution of the wine and semiquantitative analysis (without external calibration) using In as internal standard. Thirty-one elements at concentrations ranging from 0.1 mg mL–1 to 0.5 ng mL–1 can be determined by ICP–MS analysis with and without digestion. It was investigated whether a matrix effect observed for EtOH in the wine matrix can be overcome by application of a micro-concentric nebulizer with a membrane desolvator (MCN 6000). The results obtained for the MCN 6000 are compared with those obtained by use of a conventional Meinhard nebulizer. It is shown that the observed matrix effect can only be compensated by use of an internal standard for the Meinhard nebulizer, but not for the MCN 6000. Results for ICP–MS are compared with those obtained by total reflection X-ray fluorescence spectrometry (TXRF).
Introduction There is substantial interest in trace and ultratrace element determinations in the matrix wine and many articles can be found in the analytical literature, including investigations of adulteration, studies of changes in the concentration of the elements during vinification, and studies of their toxicological or physiological effects. Increasing numbers of element patterns are investigated for provenance testing.
Dedicated to Dr. habil. Hans-Joachin Dietze on the occasion of his 65th birthday M. M. Castiñeira () · R. Brandt · A. von Bohlen · N. Jakubowski Institut für Spektrochemie und Angewandte Spektroskopie (ISAS), P.O. Box 10 13 52, 44013 Dortmund, Germany e-mail:
[email protected]
Atomic spectrometric techniques, e.g. those based on inductively coupled plasma atomic emission spectrometry (ICP–AES) [1–4], mass spectrometry (ICP–MS) [5–10], neutron activation analysis (NAA) [5], or total reflection X-ray fluorescence (TXRF) [11–13] have been most commonly used to perform multi-element determination in wine. ICP–MS is preferred for routine analysis because it affords higher selectivity and sensitivity (lower limit of detection) than ICP–AES, and although NAA is a technique widely used in reference-material certification, it is limited by serious interferences which arise from major activation products. NAA has, nevertheless, been successfully used for comparison with ICP–MS in multi-elemental analysis of wines by Pérez-Jordán et al. [5]. TXRF [12, 14] is also a simultaneous multi-element technique, but it is rarely used for wine analysis. All elements with atomic numbers greater than 14 (Si) can, theoretically, be determined in the ng mL–1 range. Lighter elements can be detected only if special instrumentation is employed with a vacuum chamber, an X-ray tube with low-energy radiation, and a detector with an extremely thin window. Haswell and Walmsley [11] have already provided TXRF data for Br, Ca, Cr, Cu, Fe, K, Mn, Rb, Sr, and Zn with < 10% RSD by pipetting 5 µL samples directly on to a quartz glass reflector then adding vanadium solution (10 µg mL–1, 5 µL) as internal standard. Most application of ICP–MS in wine analysis has been devoted to the determination of rare earth elements; there has been little study of multi-element composition. Most analysis has been quantitative, but semi-quantitative approaches based on internal calibration have also been developed, for instance by Pérez-Jordán et al. [5] who used a cross-flow nebulizer, diluted the sample 1 + 1, and used Rh (10 µg mL–1) as internal standard. Wine typically contains 8–12% EtOH which can cause additional spectral and non-spectral interference in ICP–MS. To overcome matrix-induced effects, which can change the signal intensity, some authors have simply diluted the sample 1 + 1 [5, 7] or 1 + 2 [10] or removed the alcohol from the sample by evaporation to dryness and
554
dissolution of the residue in nitric acid [8, 9]. Others have applied microwave-assisted acid digestion of the wine [6]. UV-irradiation assisted digestion has also been used for sample preparation [15]. The accuracy of ICP–MS determinations has been evaluated by inter-comparison using different nebulization systems, by recovery studies, or by the analysis of reference materials different from wines. Because of spectral interference ICP-sector field mass spectrometers have usually been used – in high-mass-resolution mode. The main objective of the work reported in this paper was the development of a procedure suitable for routine direct determination of trace elements in wines by ICP–MS for provenance testing. To overcome ICP–MS matrix effects caused by EtOH, two nebulization systems were compared – the traditional Meinhard nebulizer and the micro-concentric nebulizer (MCN-6000, CETAC Technologies, Omaha, USA), which incorporates a desolvation membrane to remove the organic solvent from the aerosol. This can be used to eliminate spectral and non-spectral interferences in ICP–MS and was investigated to determine whether it could be used to overcome the matrix effect of EtOH. Two different digestion procedures were also evaluated and we also applied TXRF for validation, because no matrix effects are observed in the use of this technique, sample preparation is much more easier, and simple quantification is possible by use of well known relative sensitivities [12, 13]. The results obtained by ICP–MS and TXRF are compared and discussed.
Experimental Samples and sample preparation A red test wine (year of production 1978) with a 9.3% (v/v) EtOH content, provided by Prof. Eschnauer (Institut für Önologie, Ingelheim, Germany), was analyzed. The sample was always filtered through a 0.45-µm pore size cellulose filter to avoid clogging of the MCN. For direct analysis filtered wine (5 mL), HNO3 (subboiled p.a. quality, Merck, Darmstadt, Germany; 100 µL), and a solution of In (10 µg mL–1, 20 µL) as internal standard (stock solution 2000 µg mL–1 prepared from Fixanal, Riedel de Haën, Seelze, Germany) were mixed and diluted 1 + 1 with twice-distilled H2O. For digestion of the test wine two techniques were employed: 1. Microwave digestion (Multiwave, Perkin-Elmer, Überlingen, Germany) of wine (2.5 mL) and HNO3 (subboiled p.a. quality, Merck; 2.5 mL) in a closed system. A power of 1000 W and a temperature of 190 °C (max) were applied for 40 min (including 15 min cooling). 2. High pressure ashing (HPA, Kürner, Rosenheim, Germany) of wine (2 mL) and HNO3 (subboiled p.a. quality, Merck; 3 mL). A temperature of 280 °C resulting in a pressure of 13 MPa (max.) was applied for 60 min. Instruments ICP–MS. The ICP–MS used in this investigation was a VG Plasma Quad PQ2 Turbo Plus (Thermo Elemental, Winsford, Cheshire, UK). For sample introduction two different nebulizers were used: 1. the traditional Meinhard nebulizer (Type A, Meinhard, Santa Ana, CA, USA) with a Scott spray chamber cooled at 5 °C; and 2. the MCN 6000 (CETAC Technologies, Omaha, USA). Both nebulizers were supplied by means of a peristaltic pump (Minipuls 3, Gilson, Middleton, CT, USA). The MCN 6000 consists of a micro-concentric nebulizer, a spray chamber heated to 70 °C, and a micro-porous membrane desolvator. Heating of this
Table 1 Different ICP–MS operating and acquisition conditions used for direct semiquantitative analysis of wine by use of a Meinhard/Scott and a MCN-6000 nebulization systems
ICP
Forward power Reflected power
1400 W 5W
1350 W 8W
Gas flow rates (L min–1)
Cool gas Auxiliary Nebulizer
15 0.5 0.85
14 2.8 0.9
Nebulization system
Nebulizer Spray chamber
MCN 6000 PTFE chamber heated at 70 °C
Peristaltic pump Sample uptake rate
Meinhard, Type A Water-cooled double-pass spray chamber (quartz) at 5 °C Gilson Minipuls 3 0.4 mL min–1
Gilson Minipuls 3 0.05 mL min–1
Mass range Mode Channels per amu Dwell time (PC) Dwell time (Analogue) Number of sweeps Time per sweep Detection mode Acquisition time
5.6–239.4 amu Scanning 19 320 µs 640 µs 47 0.5 s Dual 60 s
5.6–239.4 amu Scanning 19 320 µs 640 µs 47 0.5 s Dual 60 s
Optimized on 140Ce in 5% (v/v) EtOH
140Ce
Acquisition settings
Ion lens settings
Optimized on in 1% HNO3
555 membrane to 160 °C was used to remove volatile components of the sample, e.g. EtOH, by means of a counter flow of Ar (sweep gas). The dried aerosol was then transported to the ICP. The experimental conditions employed with both nebulizers are summarized in Table 1. Details of the ICP–MS measurements, including optimization of the instrument and the quantification using the semi-quantitative approach have been described elsewhere [6]. TXRF. Simultaneous element determination in wine samples was performed by use of a TXRF-spectrometer Extra II (Rich. Seifert, Ahrensburg, Germany) and a QX 2000 detector/analyzer system (Link Systems, Oxford Instruments, High Wycombe, UK). The primary X-rays were generated by line-focus tubes with a Mo and W anode and were operated at 50 kV and 38 mA. The acquisition time for each spectrum was adjusted to 300 s. For sample preparation HNO3 (subboiled; 90 µL) and Se standard solution (1000 µg mL–1 in 5% HNO3; Johnson Matthey, Alfa Products, Karlsruhe, Germany; 10 µL) were added to the test wine (900 µL). The spiked sample solution (20 µL only) was dried on a siliconized quartz-glass carrier before analysis.
A
Results and discussion B
As mentioned previously, wine is a complex matrix that contains many inorganic and organic substances which can affect signal intensity in ICP–MS. A matrix effect arising from organic solvents, e.g. EtOH, present at elevated levels is well known in the literature but has been investigated in detail for the wine matrix by Augagneur et al. only [8]. We have, therefore, investigated this problem in more detail firstly by application of an appropriate nebulization system and secondly by digestion of the wine sample. The study described in the first section is a prerequisite for application of a semi-quantitative approach for the quantification of as many elements as possible in this matrix, which is applied in the second section.
Matrix effects of EtOH and wine To study the matrix effect caused by increasing the EtOH content (1–10% v/v EtOH) two different nebulization systems (MCN 6000 and Meinhard) were investigated using a multi-element solution containing 20 ng mL–1 Be, Mg, Co, Ni, In, Ce, Pb, and Bi in 0.14 mol L–1 HNO3 (Standard solution of 10 µg mL–1 in 2% HNO3 m/v; Spex Industries, Gasbrunn, Germany). All blanks and standards solutions were prepared in twice-distilled water and EtOH (p.a. quality, Merck, Darmstadt, Germany). The elements chosen cover almost the whole mass range and have different ionization energies, and so are representative of others, also. Fig. 1 A shows the signal intensity of the chosen elements as a function of EtOH content for the Meinhard nebulizer. It is apparent from this figure that small amounts of EtOH (approx. 0.5%) have a positive effect, increasing the sensitivity toward elements such as Co, Ni, or Be, whereas signal depression is apparent for higher concentrations up to 10%. This variation of the signal intensity can to some extent be compensated by normalization to the intensity of In, the internal standard (Fig. 1 B).
Fig. 1 A Effect of EtOH content on the ICP–MS signal from a 20 ng mL–1 multi-element solution. B Correction of the ICP–MS signal by internal standardization with In (20 ng mL–1). The Meinhard nebulizer was used
If this is done the response is more or less constant up to 4–5% EtOH, but above this concentration the signals for Ce and Co can no longer be corrected and wines with ethanol concentrations above 10% cannot be analysed directly. These measurements do, however, show that 1+1 dilution of a wine is sufficient for this nebulizer to overcome the matrix effect if normalization to In the internal standard is applied. With the MCN (MCN 6000) completely different behavior is observed for all the elements, as is apparent from Fig. 2 A. With this nebulizer the effect of EtOH on the signal intensity is extremely low; this can be attributed to complete removal of the alcohol, and if In is used as internal standard this slight effect on signal intensity is completely corrected (Fig. 2 B). From this standpoint this nebulizer is more suitable for our purpose than the traditional Meinhard nebulizer. The behavior of both nebulizers was then investigated for a wine matrix containing not only EtOH but also a variety of other organic and inorganic components. For this experiment a multi-element standard solution with each element at a concentration of 20 ng mL–1 was acidified to 0.14 mol L–1 HNO3 and used as spike for the 1 + 1 diluted wine test-sample. Recovery results obtained for 1 + 1 diluted test wine are shown in Fig. 3 A for the Meinhard nebulizer and in Fig. 3 B for the MCN 6000, for a solution containing a multielement standard solution at a concentration of 20 ng mL–1. The signal suppression (white bars) from wine components except EtOH is less important for the Meinhard nebulizer whereas it is severe for the MCN 6000. Although the MCN was more sensitive than the Meinhard nebulizer for the EtOH matrix, it was less sensitive for the
556
A A
B Fig. 2 A Effect of EtOH content on the ICP–MS signal from a 20 ng mL–1 multi-element solution. B Correction of the ICP–MS signal by internal standardization using In (20 ng mL–1). The MCN 6000 nebulizer was used
wine matrix. Correction by use of In works well for the Meinhard nebulizer and is satisfactory for the MCN 6000, except for lead. A matrix effect during use of the MCN 6000 has also been observed by Martin and Volmer [16]. They studied the effect of different matrix concentrations on signal intensity for internal standard Rh at a constant concentration of 1 ng mL–1 while the concentration of the matrix elements was varied. An increase of more than 100% of the signal intensity of the Rh signal was observed for the MCN 6000 when the concentration of the matrix components was increased from 0 to 6 ng mL–1 (concentration per element; the multi-element standards consisted of 30 elements). Because of the pronounced matrix effects and the significantly reduced sensitivity of the MCN 6000 we therefore decided to choose the Meinhard nebulizer for all further investigations. Comparison of results obtained by ICP–MS and TXRF For rapid analysis and a high sample throughput the procedure of a semi-quantitative analysis is advantageous. With this technique predetermined sensitivity factors are used and a response curve is calculated (details are given elsewhere [17]). For the VG instrument such a response curve can be generated by the software by conducting one measurement on a multi-element standard solution. This procedure can lead to rough estimates of true concentra-
B Fig. 3 A Recovery of a multi-element standard solution from 1 + 1 diluted test wine. The results are normalized to In as internal standard (black bars) and to a multi-element standard solution in 5% EtOH (white bars). The Meinhard nebulizer was used. B Recovery of a multi-element standard solution from 1 + 1 diluted test wine. The results are normalized to In as internal standard (black bars) and to a multi-element standard solution in 5% EtOH (white bars). The MCN 6000 nebulizer was used
tion values with deviations of up to a factor of 2. The response curve for this instrument had been calibrated using a multi-element solution diluted with 5% EtOH. The applicability of this approach must be validated for each new matrix either by analysis of standard reference materials, which are not available for this matrix, or by use of an independent method – TXRF in this work. The element concentrations determined for a wine sample, after 1 + 1 dilution and microwave-assisted digestion, by ICP–MS in semi-quantitative mode, and by use of TXRF, are compiled in Table 2. For ICP–MS approximately 31 elements could be determined whereas for TXRF only 12 elements could be determined. Only quantitative results for Ca, Mn, Fe, Ni, Cu, Zn, Rb, Sr, and Pb can therefore be compared. These generally show sufficient (better than ± 50%) agreement between the techniques. The agreement is even better than ± 20% for the elements Mn, Fe, Rb, Sr, and Pb. The result for Zn is surprising, because with both methods the concentration determined was higher than the results after digestion shown in Table 2. This might be indicative of partial loss of Zn during the digestion procedure, possibly as insoluble Zn oxide.
557 Table 2 Concentration of elements in the test wine after microwave-assisted digestion, after HPA digestion, and measured directly (diluted 1 + 1), by ICP–MS with use of the Meinhard nebuElement
Li B Al Ca Mn Fe Ni Cu Zn As Rb Sr Cd Sn Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Pb P S K
m/z
7 10 27 43 55 57 60 65 66 75 85 88 111 120 133 138 139 140 141 146 147 151 157 159 163 165 166 169 174 175 208 – – –
lizer in the semi-quantitative approach with In as internal standard, and by quantitative TXRF with use of Se as internal standard
ICP–MS
TXRF
MW digested (n = 6)
HPA-digested (n = 3)
Direct (n = 3)
RSD (%)
Direct 1 + 1, (n = 1) (ng mL–1)
(ng mL–1)
RSD (%)
(ng mL–1)
(ng mL–1)
RSD (%)
49.7 1860 (n = 2) 2000 138000 1030 7450 60 2960 6350 5.4 346 1270 3.2 71 0.7 103 10.3 (n = 4) 17.6 (n = 4) 2 9.2 2.4 0.6 4.2 0.6 4 1 3.6 0.5 4.1 0.6 477 n.d. n.d. n.d.
6.4 – 8.8 5.1 2.7 3.2 4.2 2.5 1.4 15 2.5 1.7 12.5 6.8 10 4.4 4.8 2.3 7.5 8.7 12.5 25 24 3.3 5.2 10 11 14 4.9 17 15 – – –
38.6 1450 1560 116000 953 6690 54 2740 6010 5 313 1190 n.d. 50 0.95 109 9.8 (n = 2) 17.2 2 8.6 1.9 0.6 3.7 0.6 4.7 1.1 3.6 0.6 3.5 0.6 448 n.d. n.d. n.d.
3.9 4.9 6.3 2.0 0.9 3.8 5.7 1.2 1.3 7.8 2.5 1.1 – 16 14 21 – 0.5 0.05 1.2 0.2 0.1 0.7 0.1 0.1 0.6 0.1 0.1 0.08 0.3 0.1 – – –
38.6 1720 2200 105000 1230 7160 (Fe 56) 60 3390 (Cu63) 10900 15 259 1500 3.8 70.7 0.8 106 10.3 17.6 2 9.6 1.9 0.45 3.9 0.63 4.5 1.1 3.5 0.5 3.8 0.6 324 n.d. n.d. n.d.
n.d. n.d. n.d. 98000 1500 9600 140 6000 11200 n.d. 440 1900 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 440 18000 130000 630000
– – – 1.1 5.2 1.4 17 1.5 1.1 – 7.2 2.7 – – – – – – – – – – – – – – – – – – 8.5 9.2 2.3 0.4
n.d.: not detected
For provenance testing the nutritional elements P, S, and K are essential; here they are easily determined by TXRF. The literature shows that either ICP-OES [3] or ICP-SFMS [8], operating in high-mass-resolution mode, have been used.
Conclusions By use of ICP–MS coupled to a conventional Meinhard nebulizer, 31 elements in the concentration range from pg mL–1 to µg mL–1 can be determined directly in a fast semiquantitative approach, after simple 1 + 1 dilution, by using In as internal standard to compensate matrix effects. When the EtOH content of the original wine is approximately
11–12%, however, this might be insufficient and a further dilution might be beneficial. For direct analysis the MCN 6000 nebulization system can only overcome matrix effects caused by EtOH but is seriously limited by other organic and inorganic components of the wine matrix. Only 12 elements in wine can be directly determined by TXRF; these include elements like P, K, or S which are difficult to determine by ICP–MS. Comparison of concentrations measured for Mn, Fe, Ni, Cu, Zn, Rb, Sr, and Pb by ICP–MS and by TXRF indicate that the ICP–MS procedure proposed is sufficiently accurate. Both TXRF and ICP–MS can be used for validation of results and are complementary techniques for the analysis of wine; they will, therefore, be applied in future work for provenance testing.
558 Acknowledgments Maria del Mar Castiñeira acknowledges a grant from the Socrates program (ref. 202-DE-13–97/2) and the assistance of Dr Montes Bayón from University of Oviedo (Spain). This study was financially supported by the German Bundesministerium für Bildung und Forschung, the Ministerium für Schule, Wissenschaft und Forschung des Landes Nordrhein-Westfalen and the German Research Council (DFG, contract No. Ja 611/2–1).
References 1. Kwan W-O, Kowalski BR, Skogerboe RK (1979) J Agric Food Chem 27:1321–1326 2. Latorre MJ, Garcia- Jares C, Medina B (1994) J Agric Food Chem 42:1451–1455 3. Thiel G, Danzer K (1997) Fresenius J Anal Chem 357:553–557 4. Muranyi Z, Papp L (1998) Microchem J 60:134–142 5. Pérez- Jordán MY, Soldevila J, Salvador A, Pastor A, de la Guardia M (1999) J Anal At Spectrom 14:33–39 6. Jakubowski N, Brandt R, Stuewer D, Eschnauer HR, Görtges S (1999) Fresenius J Anal Chem 364:424–428 7. Baxter MJ, Crews HM, Dennis MJ, Goodall I, Anderson D (1997) Food Chem 60:443–450
8. Augagneur S, Medina B, Szpunar J, Lobinski R (1996) J Anal At Spectrom 11:713–721 9. Greenough JD, Longerich HP, Jackson SE (1996) Can J Appl Spectrosc 41:76–80 10. Stroh A, Brückner P, Völlkopf U (1994) At Spectrosc 15:100–106 11. Haxwell SJ, Walmsley AD (1998) J Anal At Spectrom 13:131–134 12. Klockenkämper R, von Bohlen A (1989) Spectrochim Acta B 44:461–469 13. Carvalho ML, Barreiros MA, Costa MM, Ramos MT, Marques MI (1996) X-ray Spectrom 25:29–32 14. Klockenkämper R, von Bohlen A (1996) X-ray Spectrom 25:156–162 15. Marisa C, Almeida R, Teresa M, Vasconcelos SD (1999) J Anal At Spectrom 14:1815–1821 16. Martin M, Volmer DA (1999) Rapid Commun Mass Spectrom 13:84–86 17. J. Olesik, Inorganic Mass Spectrometry, ed. by C. M. Barshik, D. C. Duckworth, D. H. Smith, Marcel Dekker, Inc., New York, 2000, p. 122
