Determination of lanthanides by source excited energy dispersive X-ray fluorescence (EDXRF) method after preconcentration with ammonium pyrrolidine dithiocarbamate (APDC)

Analytica Chimica Acta 570 (2006) 277–282

Determination of lanthanides by source excited energy dispersive X-ray fluorescence (EDXRF) method after preconcentration with ammonium pyrrolidine dithiocarbamate (APDC) Visnja Orescanin ∗ , Luka Mikelic, Vibor Roje, Stipe Lulic Rudjer Boskovic Institute, Centre for Marine and Environment Research, Bijenicka cesta 54, 10002 Zagreb, Croatia Received 20 December 2005; received in revised form 6 April 2006; accepted 11 April 2006 Available online 27 April 2006

Abstract A new analytical procedure for determination of lanthanides in environmental samples after chemical separation from major matrix elements on DOWEX 50W-X8 resin followed by preconcentration with chelating agent ammonium pyrrolidine dithiocarbamate (APDC) and analyses of thin targets by energy dispersive X-ray fluorescence (EDXRF) method using 109 Cd as the source of excitation was presented. Characteristic L␣ X-ray lines of the lanthanides were used for calculations of the net peak area and mass concentrations. The influence of pH value of the solution and addition of organic matter on the complexation was investigated. Percentage of recovery of each lanthanide after separation on DOWEX 50W-X8 resin was also determined. Accuracy of the method was tested on standard reference materials and real environmental samples (red mud material). For that purpose samples of standard reference materials and red mud were prepared as thick targets and directly analyzed (without the separation step) by EDXRF method using 241 Am as the excitation source. In that case lanthanides concentrations were determined over their characteristic K␣ X-ray lines and results were compared with those obtained after separation/preconcentration step described above. Results showed that selected lanthanides made stable complexes with APDC in the alkaline medium with the maximum recovery at pH = 8. The presence of organic matter slightly modified the complexation by means of somewhat higher recovery percentage at pH lower than 7 and approx. 20% lower recovery at pH higher than 7. Recovery of the elements after separation on DOWEX 50W-X8 resin and preconcentration with APDC at pH = 8 varied from 91.4% (Pr) to only 24.9% in the case of Dy. Concentrations of lanthanides measured in standard reference material and environmental samples of red mud after microwave digestion, separation on DOWEX 50W-X8 resin, preconcentration with APDC at pH = 8 and recalculation on the percentage of recovery were in good agreement with certified values in the case of SRM as well as with the concentrations obtained by direct determination over K␣ lines using 241 Am excitation source in the case of red mud leading to the conclusion that presented method was applicable for the determination of lanthanides in real environmental samples. © 2006 Elsevier B.V. All rights reserved. Keywords: Lanthanides; Preconcentration; APDC; EDXRF

1. Introduction Lanthanides are used in industry due to their metallurgical, optical and electronic properties. Also, they are constantly being used in the nuclear power and nuclear weapon industries for both practical and experimental purpose. Many analytical techniques are used for determination of rare earth elements in various sample matrices. Among them the

∗

Corresponding author. Tel.: +385 1 4560934; fax: +385 1 4680205. E-mail address: [email protected] (V. Orescanin).

0003-2670/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2006.04.028

most common are: inductively coupled plasma mass spectrometry (ICP-MS) [1–5], instrumental neutron activation analysis (INAA) [6] and liquid chromatography (LC) [7]. Only few papers presented analysis of lanthanides using energy dispersive X-ray fluorescence method (EDXRF) [8]. Overlapping of the characteristic L X-ray lines of the lanthanides present in the environmental samples in relatively low concentrations with respective K X-ray lines of more abundant elements like Ti, V, Cr, Mn, Fe represents the major problem in determination of lanthanides with tube excited EDXRF method. In order to avoid interferences of major elements, different methods have been developed to separate matrix elements from

278

V. Orescanin et al. / Analytica Chimica Acta 570 (2006) 277–282

lanthanides [5,7,9–22]. Among them separation of lanthanides on ion-exchange resins was the simplest one. The columns are loaded with week acid solutions containing lanthanides, major and trace elements. Matrix elements are eluted by rinsing columns with the solution of higher acidity followed by elution of the lanthanides with strong acids. The concentrations of lanthanides in environmental samples could also be determined over their characteristic K X-ray lines without any pre-treatment. Sample preparation consists only of milling, homogenization, pressing into pellets and irradiation with 241 Am excitation source. Although highly efficient in the analyses of high Z elements, this method was not accepted in the majority of XRF laboratories worldwide due to some disadvantages in the analyses of low and medium Z elements among which are: appearance of Pb and In from the protective shield of the source in the spectrum, as well as appearance of characteristic L lines of Np as a consequence of electron capture decay of Am, poor efficiency for low Z elements, prolonged measurement time, special needs for radioactive material handling. On the other hand, majority of the XRF laboratories used standard 30 mm2 Si(Li) detectors or even thinner Si-PIN thermoelectrically cooled photodiode (7 mm2 ) and silicon-drift detectors with Mo or Rh X-ray tube in secondary target geometry or with direct (broadband) excitation. Such systems were highly efficient for the analyses of elements with the energies of their characteristic K and L lines lower than 15 keV while above 15 keV efficiency of such systems decreased to less than 20%. Since the energies of characteristic K lines of lanthanides are above 33 keV there was no possibility to analyze them over K lines with the systems mentioned above. The only possibility with such systems is the analysis of lanthanides over L lines, which lie in the energy range of appropriate system efficiency. In that case, as stated above, dissolution and separation from the major elements was necessary. Due to a relatively low sensitivity of the XRF method, direct analysis of the elements from the liquids is not applicable and some preconcentration procedure with chelating agent and preparation as thin target was mandatory. In the presented study our intention was to develop a suitable method for the above-mentioned XRF users and to give them the opportunity to expand their researches to lanthanide elements that can offer useful information in the investigations of different geochemical environments [4,6,10,11]. This paper deals with the development of analytical procedure for determination of lanthanides in iron rich matrices after chemical separation from the major matrix elements on DOWEX 50W-X8 resin [8] followed by preconcentration of lanthanides with chelating agent ammonium pyrrolidine dithiocarbamate (APDC) and analyses by EDXRF method as thin targets. 2. Experimental 2.1. Reagents All solutions were prepared using analytical reagents grade chemicals and distilled Milli-Q water. Lanthanides stock solution (concentration of each element 1 mg/L) was prepared from MERCK 1000 mg/L standard solu-

tions of each element (La(NO3 )3 , Ce(NO3 )3, Pr2 O3 , Nd2 O3 , Sm2 O3 , Gd2 O3 and Dy2 O3 dissolved in 3% HNO3 ). APDC solution was prepared daily by dissolving APDC (Aldrich) in distilled water to produce 1% (w/v) solution. One milligram per liter solution of disodium-EDTA (titriplex III) was prepared from KOMPLEKSAL II (Kemika). 2.2. Preconcentration of lanthanides with APDC One hundred milliliter of solution containing 1 mg/L of each lanthanide (La, Ce, Pr, Nd, Sm, Gd, Dy) was adjusted to pH values 3–11 by the addition of hydrochloric acid and ammonium hydroxide in order to estimate the effect of pH on the recovery of all lanthanides. All pH measurements were made with a Mettler Toledo digital pH meter. The influence of organic matter on the recovery of each element over the whole pH range was also tested. This was carried out to simulate the conditions occurring in natural materials. For that purpose lanthanide stock solution (1 mg/L of each element) was mixed with 1 mg/L of titriplex III. After the pH adjustment 2 mL of 1% (w/v) APDC was added into each flask. After the complexation lasted for 20 min, the suspension was filtered through a Millipore HAWP filter (pore size: 0.45 ␮m; diameter: 25 mm). A Millipore micro filtration system was used for that purpose. Prepared thin targets were air dried, protected by thin mylar foil (2 ␮m), inserted into a plastic carrier and placed 0.5 mm above the X-ray source of the X-ray spectrometer. 2.3. Determination of the net peak area of characteristic X-rays from the samples All targets were analyzed by energy dispersive X-ray fluorescence (EDXRF). Samples were irradiated by X-rays generated from the 109 Cd annular source. The incident angle was 49.76◦ . Detection of characteristic X-ray radiation from the sample was conducted with a Si(Li) detector (Canberra) cooled with liquid nitrogen with the following characteristics: detector size = 30 mm2 , Si thickness = 3 mm, Be window = 0.025 mm, FWHM for 5.9 keV 55Fe 165 eV. The emerging angle was 74.05◦ and the distance was 1.5 cm. Spectra were collected by Genie – 2000 software (Canberra, Meriden, CT USA). Spectral data were analyzed by WinAxil software version 4.5.2 (Canberra Eurisys Benelux, Belgium) using characteristic L␣ lines of the elements. In order to obtain a good counting statistic, collecting time for all targets was 10,000 s. Relative recovery was calculated by dividing the net peak area (N) of characteristic L␣ X-ray lines obtained for each lanthanide in each target with WinAxil software by maximum net peak area (N0 ) which was obtained from the target prepared at pH 8. Relative recovery of lanthanides in the presence of EDTA was calculated by dividing the obtained net peak area (NEDTA ) with N0 . Calibration file (model) for the quantitative analyses was created on the basis of the measurements of the standard solution (Merck) containing 1 mg/L of La, Ce, Pr, Nd, Sm, Gd, Dy preconcentrated with APDC at pH = 8 which was found as an optimum pH value for the chelation of the majority of the tested lanthanides. Quantitative analysis was done by the “Compared

V. Orescanin et al. / Analytica Chimica Acta 570 (2006) 277–282

method” from the WinFund package, version 4.5.2 (Canberra Eurisys Benelux, Belgium). Preparation of samples as thin targets was the reason for setting attenuation of X-ray lines in the sample and enhancement effect to 1 during the calculation of calibration constant. All other parameters like the detector efficiency, fundamental parameters, intensity distributions of the characteristic lines and the continuum, a factor taking into account the overall intensity of the source and the measurement time were taken into account. Minimum detection limit (MDL) for each unit lanthanide concentration (1 mg/L) and measuring time of 10,000 s were determined according to Van Grieken & Markowicz [23]. 2.4. Recovery of lanthanides after separation on ion-exchange resin The influence of high concentrations of iron, manganese and zinc, which were present in most of the environmental samples was investigated by adding 5 mg/L of these elements in stock solution containing 1 mg/L of La, Ce, Pr, Nd, Sm, Gd, Dy. The sample solution was passed through the column with DOWEX 50W-X8 (100–200 mesh) cation resin pre-equilibrated with 1 M HCl. Matrix macro elements were eluted with 60 mL of 1.7 M HCl. The lanthanides were eluted with 45 mL of 6 M HCl [8]. Obtained solution of lanthanides was evaporated almost to dryness, diluted to 100 mL, preconcetrated with 2 mL of APDC at pH 8 as described above and analyzed. The same procedure was repeated six times and the results were expressed as mean values of these six measurements. Obtained mean values were divided with initial concentrations and multiplied with 100 in order to calculate the percentage of the recovery of each lanthanide. 2.5. Analyses of lanthanides in environmental samples as thin targets 0.123 g of USGS standard G-2 and powdered red mud were digested with 10 mL of nitric acid (p.a., Kemika Zagreb) by means of the microwave sample preparation system Anton Paar Multiwave 3000. The digestion conditions applied for the microwave system were: tmax = 230 ◦ C, 20 min ramp, 20 min at 230 ◦ C and 20 min vent. After cooling, the solutions were diluted to 100 mL by Milli-Q water. Lanthanides from the sample solutions were separated with DOWEX 50W-X8 resin as described above, preconcetrated with APDC at pH = 8 and analyzed. Six replicate samples of each material were analyzed and expressed as mean value of these six measurements. 2.6. Direct XRF analyses of lanthanides in environmental samples using 241 Am source In order to demonstrate the accuracy of the developed method 2 g of different standard reference materials like USGS G-2; BCR-2; AGV-1 as well as IAEA SL-1 and 8 samples of red mud were pressed into pellets and irradiated with 241 Am annular source for 10,000 s. Calibration files (models) for spectrum fitting and quantitative analyses were created on the basis of the

279

measurement of USGS G-2 standard reference material using “Fundamental parameters” method from the WinFund package version 4.5.2 (Canberra Eurisys Benelux, Belgium). Characteristic K␣ lines of the elements were used in the model. The concentration (wi ) of each lanthanide (i) was calculated according to the following formula: wi = Nij /(Kij Dij FPij Aij,m Hij,m Ii0 t) where Nij is the net peak area of the characteristic X-ray line j; Kij the calibration constant calculated on the basis of the standard reference material; Dij the detector efficiency at Eij energy; FPij the number of fundamental parameters (like ionization cross section, fluorescence yields. . .); Aij,m the attenuation of the X-rays due to matrix (m); Hij,m the enhancement of the X-rays due to the matrix (m); I the intensity distribution of the characteristic lines and of the continuum in the excitation; i0 the factor taking into account the overall intensity of the source (the activity of the radio isotope); t the measurement time. Accuracy of the model was tested on the other three standard reference materials. After obtaining satisfying results on SRMs, the model was used for determination of the lanthanides in the red mud samples.

3. Results and discussion 3.1. The influence of pH and organic matter on the recovery of lanthanides The effect of pH’s between 3 and 11 on the relative recovery of each lanthanide in the absence of other substances and in presence of organic matter was presented in Fig. 1. The elements Pr, Nd, Sm, Gd and Dy showed very similar recovery patterns with more or less linear increase from pH 3, with no or minimum recovery (approx. 1%), to pH 6 (with approx. 40% relative recovery). At pH 7 complexation of all mentioned elements with APDC decreased significantly while the optimum results were obtained at pH = 8 and decreased slightly toward pH 11. The presence of organic matter slightly modified the complexation by means of somewhat higher recovery percentage at pH lower than 7 and approx. 20% lower recovery at pH higher than 7. La and Ce showed more irregular patterns compared to other five elements. Obtained recovery for La was less than 20% at pH = 3 and 4 and slightly decreased to 0 at pH = 7, reaching maximum value at pH = 9 and slightly decreased to 70% at pH = 11. Similar pattern with somewhat lower recovery of the lanthanum was found in the presence of EDTA. In the pH range of 3–7 recovery of Ce ranged from 15 to 25%, while maximum value was obtained at pH = 8. Very high recovery for Ce were also found at pH = 9 (91%), pH = 10 (98%) and pH = 11 (83%). In the presence of EDTA recoveries for Ce continuously increased with pH value of the solution reaching maximum value at pH = 8 (80%) and further decreased from pH 9 to 11.

280

V. Orescanin et al. / Analytica Chimica Acta 570 (2006) 277–282

Fig. 1. The effect of pH between 3 and 11 on the relative recovery of each lanthanide in the absence/presence of organic matter.

3.2. Results of the recovery of lanthanides after separation on ion-exchange resin Table 1 presents the following results of the EDXRF measurement of lanthanides prepared as thin targets using 109 Cd as excitation source: characteristic L␣ lines, minimum detection limit for each element, repeatability of the method, recovery of each element from the solution containing 1 mg/L of each lanthanide and 5 mg/L of the elements Mn, Fe and Zn after separation on DOWEX 50W-X8 resin, certified and measured concentrations of lanthanides and their uncertainties in USGS G-2 standard reference material after microwave digestion, sep-

aration on DOWEX 50W-X8 resin and preconcentration with APDC at pH = 8. Fig. 2a showed a spectrum of this thin target. The best recovery was found for the elements Pr (91.4%), Nd (94.0%) and Sm (92.9%) while the least satisfying results were obtained in the case of Dy (24.9%) and Gd (54%). Seventy-four percent of La and 66.6% of Ce were recovered after separation. In order to obtain accurate concentrations in USGS G-2 SRM after separation on DOWEX resin, measured values were recalculated on the percentage of recovery of each lanthanide presented in Table 1. Obtained results after recalculation on the percentage of recovery were in agreement with the certified values of lanthanides in USGS G-2 standard reference material. The

V. Orescanin et al. / Analytica Chimica Acta 570 (2006) 277–282

281

Table 1 EDXRF measurement of lanthanides prepared as thin targets using 109 Cd as excitation source Element

L␣ line (keV)

MDL (mg/L) (N = 10)

%S.D. (N = 10)

Recovery (%)

USGS G-2 Measured

La Ce Pr Nd Sm Gd Dy

4.651 4.838 5.032 5.228 5.633 6.054 6.491

0.77 0.55 0.24 0.24 0.35 0.17 0.17

12.2 7.8 1.7 1.7 2.5 0.8 0.8

74.0 66.6 91.4 94.0 92.9 54.0 24.9

85 154 20 51 6.6 3.9 2.5

± ± ± ± ± ± ±

Certified 89 ± 8 160 ± 10 18 55 ± 6 7.2 ± 0.7 4.3 2.4 ± 0.3

4.3 5.7 1.2 0.6 0.8 0.2 0.5

Characteristic L␣ lines, minimum detection limit for each element, repeatability of the method expressed as the percentage of standard deviation and recovery of each element from the solution containing 1 mg/L of each lanthanide and 5 mg/L of the elements Mn, Fe and Zn after separation on DOWEX 50W-X8 resin expressed as mean value of six replicates, certified and measured concentrations and uncertainties (±2 S.D.) of lanthanides presented as mean values of six replicates in USGS G-2 standard reference material after microwave digestion, separation on DOWEX 50W-X8 resin and preconcentration with APDC at pH = 8.

Fig. 2. EDXRF spectra of lanthanides: (a) thin target containing 1 mg/L of lanthanides (La, Ce, Pr, Nd, Sm, Gd, Dy) and hafnium irradiated with 109 Cd annular source after separation on DOWEX resin from 5 mg/L of Mn, Fe and Zn; (b) thick target prepared by pressing 2 g of red mud into pellet irradiated with 241 Am annular source.

Table 2 EDXRF analyses of thick targets Element

La Ce Pr Nd Sm Gd Dy

K␣ line (keV)

MDL (mg/kg)

%S.D. (N = 10)

33.297 34.561 35.855 37.176 39.903 42.745 45.709

3.46 4.24 2.45 1.73 1.34 0.95 1.34

1.51 2.54 2.01 2.70 1.49 0.68 2.41

USGS G-2

IAEA SL1

Certified

Measured

89 ± 8 160 ± 10 18 55 ± 6 7.2 ± 0.7 4.3 2.4 ± 0.3

86 164 18 55.0 6.9 4.3 3

± ± ± ± ± ± ±

4 6 2 1.0 0.9 0.3 0.6

USGS BCR-2

USGS AGV-1

Certified

Measured

Certified

Measured

Certified

Measured

52.6 ± 117 ± – – 9.25 ± – 7.5 ±

55 ± 3 114 ± 4 – – 9±2

25 ± 53 ± 6.8 ± 28 ± 6.7 ± 6.8 ± –

23 ± 46 ± 15 ± 29 ± 7± 6.2 ± –

38 ± 67 ± – 33 ± 5.9 ± 5± 3.6 ±

36 ± 2 65 ± 3 17 ± 4 36 ± 3 5.4 ± 0.9 0.8 11 ± 4

3.1 17

0.51 2.2

7±2

1 2 0.3 2 0.3 0.3

2 2 3 2 2 0.8

2 6 3 0.4 0.6 0.4

Characteristic K␣ lines, minimum detection limit for each element, repeatability of the method expressed as the percentage of standard deviation on the basis of 10 measurements of USGS G-2 standard reference material, certified and measured concentrations of lanthanides (in mg/kg) and their uncertainties (±2 S.D.) in standard reference materials prepared as thick targets, irradiated with 241 Am source and determined over K␣ lines.

282

V. Orescanin et al. / Analytica Chimica Acta 570 (2006) 277–282

Table 3 Mean values and uncertainties of lanthanides in eight samples of red mud prepared as thick targets, irradiated with 241 Am source and determined over K␣ lines, and mean values of lanthanides in six samples of red mud after microwave digestion, separation of lanthanides on DOWEX 50W-X8 resin, preparation for the analyses as thin targets, irradiation with 109 Cd source and determination over L␣ lines Element (mg/kg)

Thick target-direct determination

La Ce Pr Nd Sm Gd Dy

98.5 248.9 32.3 118.3 17.1 7.3 12.3

± ± ± ± ± ± ±

1.2 3.1 0.6 1.3 0.6 0.3 0.8

Thin target-after separation 97 267 22 103 27 8 8.8

± ± ± ± ± ± ±

4.9 9.9 1.3 1.2 3.1 0.4 1.8

In order to obtain real concentrations in red mud comparable with direct measurements, determined values for thin targets were recalculated on the percentage of recovery for each lanthanide presented in fourth column of Table 1.

differences between certified and measured values were less than 10%. 3.3. Results of direct XRF analyses of lanthanides in environmental samples using 241 Am source Table 2 presents the results of EDXRF analyses of thick targets of four standard reference materials irradiated with 241 Am source and determined over K␣ lines. Compared to the certified values satisfying results were obtained for the concentrations of all measured lanthanides in all four standard reference materials. After obtaining satisfying results on SRMs, the method was used for direct determination of lanthanides in red mud samples (Fig. 2b). Results were presented in Table 3 and compared to mean values obtained in red mud after microwave digestion, separation on DOWEX 50W-X8 resin, preconcentration with APDC at pH = 8 and recalculation on the percentage of recovery of each lanthanide. Values obtained after separation were in good agreement with the results obtained by direct measurements confirming the applicability of developed method on real environmental samples. 4. Conclusion Results showed that the selected lanthanides made stable complexes with APDC in the basic medium with the maximum recovery at pH = 8. The exception was La which reached maximum recovery at pH = 9. At pH values ranging from 9 to 11

recoveries varied from 98 to 70%. The complexation with APDC was irregular at all other pH values. The presence of organic matter slightly modified the complexation by means of somewhat higher recovery percentage at pH lower than 7 and approx. 20% lower recovery at pH higher than 7. Recovery of the elements after separation on DOWEX 50WX8 resin and preconcentration with APDC at pH = 8 varied from 91.4% (Pr) to only 24.9% in the case of Dy. Concentrations of lanthanides measured in standard reference material and environmental samples of red mud after microwave digestion, separation on DOWEX 50W-X8 resin, preconcentration with APDC at pH = 8 and recalculation on the percentage of recovery were in good agreement with certified values in the case of SRM as well as with the concentrations obtained by direct determination over K␣ lines in the case of red mud leading to the conclusion that presented method was applicable for determination of lanthanides in real environmental samples. References [1] M. Ochsenk¨uhn, Th. Lyberopulu, G. Parissakis, Anal. Chim. Acta 296 (1994) 305. [2] F.E. Lichte, A.L. Meier, J.G. Crock, Anal. Chem. 59 (1987) 1150. [3] M. Ochsenk¨uhn-Petropulu, M. Ochsenk¨uhn, J. Luck, Spectrochim. Acta 46 (1991) 51. [4] K.E. Jarvis, Chem. Geol. 68 (1988) 31. [5] T. Pasinnli, A.E. Ero˘glu, T. Shahwan, Anal. Chim. Acta 547 (2005) 42. [6] P. Vucotic, J. Radioanal. Chem. 78 (1983) 105. [7] C. Na, T. Nakano, K. Tazawa, M. Sakagawa, T. Ito, Chem. Geol. 123 (1995) 225. [8] R. Djingova, J. Ivanova, Talanta 57 (2002) 821. [9] I.E. De Vito, A.N. Masi, R.A. Olsina, Talanta 49 (1999) 925. [10] B. B¨uhn, A.H. Rankin, Geochim. Cosmochim. Acta 63 (1999) 3781. [11] L.M. Larsen, J.G. Fitton, A.K. Pedersen, Lithos 71 (2003) 47. [12] R. Padilla, P. Van Espen, P.P. Godo Torres, Anal. Chim. Acta 558 (2006) 283. [13] T. Arai, Y. Wei, M. Kumagai, K. Horiguchi, J. Alloys Compd. 408 (2006) 1008. [14] T. Suzuki, K. Itoh, A. Ikeda, M. Aida, M. Ozawa, Y. Fujii, J. Alloys Compd. 408 (2006) 1013. [15] Y. Sun, J. Chromatogr. A 1048 (2004) 245. [16] D. Li, Y. Zuo, S. Meng, J. Alloys Compd. 374 (2004) 431. [17] H. Minowa, M. Ebihara, Anal. Chim. Acta 498 (2003) 25. [18] B.A. Haley, G.P. Klinkhammer, Mar. Chem. 82 (2003) 197. [19] R. Garcia-Valls, A. Hrdlicka, J. Perutka, J. Havel, N.V. Deorkar, L.L. Tavlarides, M. Mu˜noz, M. Valiente, Anal. Chim. Acta 439 (2001) 247. [20] B. Liu, L. Liu, J. Cheng, Anal. Chim. Acta 358 (1998) 157. [21] C. Pin, J. Santos Zalduegui, Anal. Chim. Acta 339 (1997) 79. [22] M.C. Bruzzoniti, E. Mentasti, C. Sarzanini, M. Braglia, G. Cocito, J. Kraus, Anal. Chim. Acta 322 (1996) 49. [23] R. Van Grieken, A. Markowicz (Eds.), Handbook of X-ray Spectrometry, Methods and Techniques, Marcel Dekker, New York, 1993, pp. 181–293.