摘要:

JOURNAL O F ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 1059 Direct Solid Soil Analysis by Laser Ablation Inductively Coupled Plasma Atomic Emission Spectrometry* Lieselotte Moenke-Blankenburg and Thomas Schumann Martin-Luther-University Halle- Wittenberg Department of Chemistry Institute of Analytical and Environmental Chemistry Weinbergweg 76 D-06720 HallelSaale Germany Joachim Nolte Bodenseewerk Perkin-Elmer GmbH P.O. Box 707 767,D-88647 UberlingenlBodensee Germany Determination of heavy metals in soils by inductively coupled plasma atomic emission spectrometry (ICP- AES) usually involves the time-consuming step of preparing a solution of the solid that is then nebulized into the plasma. According to regulations digestion by aqua regia (hydrochloric acid + nitric acid 3 + 1) should be carried out although it is known that this method is incomplete for silicate soils.The problem can be eliminated by introducing the solid directly into the plasma using the laser ablation technique for sampling. Results are described for a study of laser ablation using a Q-switched Nd:YAG laser coupled with a new echelle spectrometer which has a multichannel solid-state detector. The laser pulses were focused onto the solid surface of pressed soil samples to generate an aerosol which is entrained in a flowing Ar stream transported through a tube and then introduced directly into the inductively coupled plasma. Some character- istics of the preparation technique the selection of an internal standard and homogeneity tests of the elemental distribution are reported along with a comparison and evaluation of three methods of calibration.The criteria used to measure the performance of laser ablation ICP-AES are the relative standard deviations obtained of 4.9-12.7% and the accuracy O.3-12A0/o for Fe Mn Cu Pb Cr Zn and Ni. Keywords Laser ablation inductively coupled plasma atomic emission spectrometry; soil analysis; heavy metals in soils; solids analysis; calibration The determination of heavy metals in soils is carried out for various reasons including the measurement of the total elemen- tal content. This provides base-line knowledge of the compo- nents in the soil with respect to which changes in soil composition produced by elution pollution plant uptake or agriculture manipulation can be assessed.' Analysis techniques can be broadly categorized into single- element methods such as atomic absorption spectrometry (AAS) or simultaneous multi-element methods such as induc- tively coupled plasma atomic emission spectrometry (ICP- AES) or X-ray fluorescence spectrometry (XRFS).They can be categorized further into methods such as AAS and ICP- AES where the analysis is carried out in solution i.e. with dissolved samples or methods such as XRFS and now laser ablation-ICP- AES (LA-ICP-AES) and LA-ICP mass spec- trometry (LA-ICP-MS) where solid samples are analysed directly. The choice of method for a particular application should take account of these factors as well as their precision and accuracy. However the choice could be dominated by the techniques currently in use.Direct solid sampling using the interaction of focused laser radiation with a sample is a potentially useful procedure in tandem techniques such as LA-ICP-AES and LA-ICP-MS.2,3 Although LA-ICP-AES has been used for a range of appli- cations only a few of these have been in the area of soil a n a l y ~ i s . ~ - ~ Therefore further investigations are still required before the method can be recommended for routine application to direct solid soil analysis. Direct solid soil analysis is of interest because standardized and validated methods (e.g. in Germany) for the analysis of sludge and sediments7 include the digestion with aqua regia (hydrochloric acid + nitric acid 3 + 1 ) for subsequent determi- nation of the heavy metals by flame AAS,' electrothermal vaporization AAS9 and solution nebulization ICP-AES." The digestion of soils with aqua regia is incomplete if silicates and acid-insoluble oxides (e.g.heated chromium and iron oxides7) * Presented at the 1994 Winter Conference on Plasma Spectro- chemistry San Diego CA USA January 10-15 1994. are present. Therefore the proportion of soluble heavy metals differs from soil to soil and the results are not comparable. Other non-standardized methods for digestion of soils require the use of hazardous reagents such as HF. In addition wet- chemical pre-treatment processes are time consuming and lead to dilution of the sample. X-ray fluorescence is a popular technique for the direct determination of major and minor elements in soils. Frequently however elements of interest are not present in abundances easily determined without significant optimization of the instrumentation. The productivity and large dynamic range of measurement of a multichannel simultaneous LA-ICP-AES instrument enables the determination of all environmentally important elements to be made and overcomes the problems mentioned above.Experimental Samples As test samples the following certified reference materials were used BCSS 1 Baie des Chaleurs Sediment Standard from the National Research Council Canada Ottawa Canada; SO 4 Reference Soil Sample from the Canada Centre for Mineral and Energy Technology Ottawa Canada; and GSS 1 3 and 8 Geochemical Soil Standards from the Chinese State Bureau of Technical Supervision in Beijing China. Sample Preparation Because the standardized and validated method of digestion is by aqua regia7 a 'behrotest ET 1 system' (Behr-Labor-Technik) was used for dissolution of the soil samples.The sample must have a grain size of less than or equal to 0.1 mm according to the recommendation^.^ The pulverized and sieved material was dried for 30 min at 105 "C in a drying oven and 3 g of the soil (weighed to within 0.01 g) were transferred into the digestion vessel. An acid mixture consisting of 7 ml of 65% nitric acid (pro analysi) with p= 1.40 gml-' (Merck) and 21 ml of 32%1060 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 hydrochloric acid (pro analysi) with p= 1.16 g ml-' (Merck) was used. The vessel was then sealed and a two-step digestion programme was performed.In step one low power to heat up to 40-50°C was used to decompose the organic material present in the sample; in step two full power up to 130°C for 2 h was used for dissolution (leaching) of the inorganic material in the soil. The resulting solution was transferred into a calibrated flask and diluted to a volume of 100ml with distilled water. After filtering the residue was dried and handled as described in the next section. As the aim of the investigation was direct solid soil analysis the residues and all of the original untreated soil reference samples with a given (certified) particle size of 6 2 pm (particle size is a critical factor in laser ablation transport tech- niques'1,'2) were subjected to a 16 h drying procedure at 105 "C. To 1 g of the dried sample 500 pl of one of five single element standard solutions (Titrisol Merck) containing 1000ppm of each of Y Be Pd Sc and In were added as internal standards.After homogenization by hand with an agate mortar pellets were pressed at 10 MPa without a binder applying a steel ring of 12mm i.d. and 3 mm thickness as a mount. The distribution of all elements of interest (Cr Cu Ni Pb Zn Mn and Fe) was tested by employing electron probe microanalysis (EPMA). In Fig. 1 is shown the distribution of Mn and Fe in soil GSS 1 after homogenization checked by the Cameback Micro from CAMECA with operating con- ditions of 20 kV and 100 nA. Instrumentation A Perkin-Elmer SCIEX laser sampler 320 was used for ablating the material.13 The sample was set in a chamber and positioned with a stepper motor in the x y and z directions.The process could be observed with a video camera. The Nd YAG laser used at a wavelength of 1064 nm was connected with a 1.5 m Tygon tube to a Perkin-Elmer Optima 3000 ICP emission spectrometer which was fitted with a dry aerosol quick change sample introduction system. Argon gas flowing through the ablation cell transported the laser ablated soil material from the ablation cell to the ICP torch. The design and operating characteristics of the LA-ICP- AES instrumentation have been described previo~sly.~ Particular operating conditions used in the present work are given in Table 1. Procedure The wavelengths suitable for LA-ICP-AES analysis of soil samples are 267.716 nm for Cr 324.754 nm for Cu 231.604 nm for Ni 220.353 nm for Pb 206.199 nm for Zn 257.610 nm for Table 1 Operating conditions ~~ Laser Wavelength Laser pulse mode Q-switch delay Laser sampling mode Laser beam energy ICP-AES R.f.generator frequency R.f. power Plasma gas flow Intermediate gas flow Aerosol carrier gas flow Observation height Read-out time Integration time ND YAG 1064 nm Q-switched 240 ps Single shot 180 mJ Echelle spectrometer with a segmented array charge coupled device detector 40 MHz free running 1100 W true power control 15 1 min-' 1 1 min-' 1 1 min-' 15 mm 50 ms 5 s H 10 pm Fig. 1 Distribution of (a) Mn and (b) Fe in soil sample GSS 1 after homogenization checked by EPMA (Cameback Micro CAMECA 20 kV 100 nA) Mn and 259.940nm for Fe. Because internal standards have to be used to compensate for pulse to pulse variations in the amount of material ablated the following elements and wave- lengths were used for the evaluation of the suitability of a particular element for internal standardization Y 371.036 nm Be 313.107nm Pd 340.458nm Sc 361.384nm and In 325.609 nm.The geological standard GSS 1 was taken as the test matrix. This matrix was mixed with standard solutions of Y Be Pd Sc and In (Merck) in appropriate concentrations prepared as described above. The relative standard deviations (RSD) were calculated after five measurements for each analyte as Cr Cu Ni Pb Zn Mn and Fe. The results are shown in Table 2 and they led to the decision to select Y which had the lowest RSD values as being suitable for the purpose of internal standardiz-h Table 2 Selection of internal standard; matrix GSS 1 RSD (Yo) Element Cr c u Ni Pb Zn Mn Fe Y 4.9 5.9 9.3 7.2 2.7 1.7 1.3 Be 12.2 3.7 25.4 12.6 1.8 0.9 3.1 Pd 19.0 13.5 5.6 2.3 4.1 1.1 0.9 s c 40.9 54.2 25.0 2.9 3.5 4.3 4.8 In t 1.3 9.4 18.7 10.8 7.5 4.0 2.7 ation in this particular case.The homogeneity of the distri- bution of Y in soil samples was checked with EPMA. The excellent result obtained is demonstrated in Fig. 2. The certified values of seven elements in soil reference materials BCSS 1 SO 4 GSS 1 GSS 3 and GSS 8 partly used as external standards for calibration or as samples for testing purposes are given in Table 3. It is commonly accepted that calibration for solids analysis should be performed with matrix-matched standards in most cases especially with such complex materials as soils.Furthermore when using external standards in laser ablation ICP spectrometry they have to be homogeneous in the micro-region (the size of the laser craters produced were about 100 pm in diameter). The standards BCSS 1 SO 4 GSS 1 GSS 3 and GSS 8 meet these require- ments and can be used as external standards in the analysis of soils. H 10 pm Fig.2 Distribution of the internal standard Y in soil sample GSS 1 after homogenization tested by EPMA (Cameback Micro CAMECA 20 kV 100 nA) Table 3 Certified values of the reference materials Element Cr/mg kg- Cu/mg kg - ' Ni/mg kg- Pb/mg kg-' Zn/mg kg - ' Mn/mg kg - ' Fe(% m/m) BCSS 1* S O 4 123 61 18.5 22 55.3 26 22.7 16 119 94 229 900 3.3 2.4 GSS 1 62 21 20.4 98 680 1760 3.7 GSS 3 32 11.4 12.2 26 31.4 304 1.4 GSS 8 68 24.3 31.5 21 68 650 3.1 Calibration with aqueous standards has been proposed by Thompson et all4 and by Moenke-Blankenburg and co-workers termed liquid-solid calibration and has been described in detai1.15-17 The concentration of the analyte element(s) in the solid can be calculated using the formula where C is the unknown concentration of the element to be determined in the solid C1 and C are the preliminary concentrations of the analyte element and the reference element using the analytical curves of the same elements obtained from the standard solutions and C is the known concentration of the reference element (internal standard) in the solid.Calibration by peak-height measurement of the transient signal of only one external standard was investigated as the third method of calibration.Using a single shot of an impulse laser the result is a spiky profile of the transient signal of a few seconds duration. The evidence of proportionality of peak heights and concentrations is demonstrated in Fig. 3 using the peak-height intensities of the Cr contents of three reference materials to plot a linear analytical curve. As a consequence one external standard with known concentrations of the elements to be determined should be sufficient for quantifi- cation of an unknown sample. In Table 4 the criteria of performances of the three cali- bration methods are compared. From the point of view of the reproducibility and the accuracy of the values the indication is that the preference is to use peak-height calibration for the investigation with only one standard.0 10 32 62 123 Concentration (pprn) Tirne/s Fig. 3 Proportionality of peak height and concentration for Cr. Reference materials BCSS 1 GSS 1 and GSS 3 Table4 calibration using GSS 1 as sample; internal standard Y; n=5 Comparison of the performances of the three methods of External standard* (YO) Liquid-solid7 (%) Peak height1 ( O h ) Element Cr c u Ni Pb Zn Mn Fe RSD Accuracy 4.9 2.2 1.9 19.0 9.3 45.1 7.3 7.2 2.8 7.5 1.7 28.1 1.4 4.5 RSD Accuracy 16.9 180.1 18.4 4.2 34.5 25.4 8.1 8.1 31.5 51.5 7.0 34.6 6.6 38.3 RSD Accuracy 3.4 9.5 10.0 1.1 9.4 3.9 3.1 0.5 2.5 0.3 4.3 1.2 5.7 2.8 * External standards BCSS 1 SO 4 GSS 3 and GSS 8. t Single element standard solutions Cr 5 10 and 20; Cu 5 10 and 2 0 Ni 5 10 and 20; Pb 5 10 and 20; Zn 5 lnd 20; 0 and 20; Mn 10 30 and 50; Fe 20 50 and 100 mg kg-'.$ External standard BCSS 1.1062 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 Table 5 Determination of heavy metals in soil SO 4 after digestion; SiO content 68.4% m/m ICP-AES of the filtrate after aqua regia digestion* Element Cr c u Ni Pb Zn Mn Fe (YO m/m) x * AX/ mg kg-' 26.1 k0.5 14.2 f 0.4 19.1 k0.4 12.3 + 0.4 91.0 f 2.0 484.2 ? 10.2 1.8f0.0 RSD ("/I 1.4 2.3 1.6 2.9 1.8 1.7 1.5 LA-ICP- AES of the residue after aqua regia digestiont X & Ax/ RSD mg kg-' (Yo) 38.3 + 3.5 7.4 7.9 f 1.0 9.8 35.4 f 6.6 14.9 17.7 f 1.5 6.6 7.8 k 1.2 12.1 113.6 & 9.5 6.7 0.6k0.1 12.9 Certified X + Ax/ mg kg-' 6156 22+ 1 26+3 16+3 94+3 600 f 20 2.37 f0.07 Found1 mg kg-' 64.4 22.1 54.5 30.0 98.8 597.8 2.4 Accuracy ("/I 5.6 0.5 109.6 87.5 5.1 0.4 1.3 * n=5.1 Sum of the contents found in the filtrate and in the residue. Internal standard Y. Peak-height calibration external standard BCSS 1 n = 5. Results and Discussion The results obtained during the investigation are presented in Table 5. The heavy metals Cr Cu Ni Pb Zn Mn and Fe were determined following the instructions in the recommended proced~re.~ After digestion with aqua regia the filtrate was analysed by ICP-AES. Because of the Si02 content (68.4%) in soil SO 4 only 42.8% of the Cr 64.5% of the Cu 73.5% of the Ni 76.9% of the Pb 96.8% of the Zn 80.7% of the Mn and 75.9% of the Fe was found. Applying LA-ICP-AES to analyse the residue remaining after digestion with aqua regia the accuracy of the difference from the certified values in the case of Cr was found to be 5.4% Cu 0.5% Zn 5.1% Mn 0.4% and Fe 1.3%.Only the results for the direct solid determinations of Ni and Pb were inadequate and could not be explained. The RSD values varied between 6.6 and 14.9% which is the range normally to be expected in LA-ICP-AES. The results for accuracy and precision were confirmed by analysing the soil sample as a whole by direct LA-ICP-AES as shown in Table 6. The precision was between 4.9 and 12.7% and the accuracy between 0.3 and 12.4%. Conclusion Direct solid soil analysis can be obtained with LA-ICP-AES after careful sample preparation providing in particular par- ticle sizes of about 2 pm in diameter and a homogeneous distribution of the elements in the pressed pellets.A comparison of three methods of calibration led to the recommendation to use peak-height measurement of the transient signal of only Table 6 Determination of heavy metals in soil SO 4 without digestion; SiO content 68.4% m/m; n = 5 Element Cr c u Ni Pb Zn Mn Fe (% m/m) X _+AX/ mg kg-' 67.0 k 4.2 23.2 2.1 29.7 k 2.7 15.25 1.8 104.6 + 16.5 6 19.4 k 40.7 2.4 k 0.1 RSD (%) 5.0 7.1 7.3 9.7 12.7 5.3 4.9 LA-ICP-AES Certified X Ax/ mg kg-' 61 +6 22+ 1 26f3 16+3 9 4 f 3 600 f 20 2.37 + 0.07 Accuracy 8.9 5.3 12.4 5.5 10.2 3.1 0.3 (%> one external standard for calibration. The criteria of precision and accuracy are of an order of magnitude of 10% and meet the desires and expectations for soil analysis. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Ure A.in Heavy Metals in Soils ed. Alloway B. J. Blackie Glasgow and John Wiley New York 1990 ch. 4 pp. 40-79. Moenke-Blankenburg L. Spectrochim. Acta Rev. 1993 15 McLeod C. W. Routh M. W. and Tikkanen M. W. in Inductively Coupled Plasmas in Analytical Atomic Spectrometry eds. Montaser A. and Golightly D. W. VCH Weinheim 2nd. edn. 1992 ch. 16 pp. 753-769. Zamzow D. S. Balswin D. P. Weeks S. J. Bajic S. J. and DSilva A. P. Environ. Sci. Technol. in the press. Nolte J. Moenke-Blankenburg L. and Schumann T. Fresenius' J. Anal. Chem. 1994 349 131. Niskavaara H. Kallio E. and Lehto O. ICP In$ Newsl. 1993 19 331. Deutsche Einheitsverfahren zur Wasser- Abwasser- und Schlammuntersuchung VCH Weinheim und Beuth Verlag Berlin 1993 DIN 38414 S7 1983. DIN 38406 E6-1 (Pb) 1981 E7-1 (Cu) 1991 E8-1 (Zn) 1980 E10-1 (Cr) 1985 Ell-1 (Ni) 1991. DIN 38406 E6-3 (Pb) 1981 E7-2 (Cu) 1991 E10-2 (Cr) 1985 Ell-2 (Ni) 1991. DIN 38406 E22 (Pb Cr Fe Cu Mn Ni Zn) 1988. Chenery S. Thompson M. and Timmins K. Anal. Proc. 1988 25 68. Arrowsmith P. and Hughes S . K. Appl. Spectrosc. 1988,42,1231. Denoyer E. R. Fredeen K. J. and Hager J. W. Anal. Chem. 1991 63,445A. Thompson M. Chenery S. and Brett L. J. Anal. At. Spectrorn. 1989 4 11. Moenke-Blankenburg L. Gackle M. Gunther D. and Kammel J. in Plasma Source Mass Spectrometry. The Proceedings of the Third Surrey Conference on Plasma Source Mass Spectrometry University of Surrey July 16th-l9th 1989 eds. Jarvis K. E. Gray A. L. Jarvis I. and Williams J. G. Royal Society of Chemistry Cambridge 1990 pp. 1-17. Moenke-Blankenburg L. Schumann T. Gunther D. KuO H.-M. and Paul M. J. Anal. At. Spectrom. 1992 7 251. Moenke-Blankenburg L. and Gunther D. Chem. Geol. 1992 95 85. pp. 1-37. Paper 4/01 360C Received March 7 1994 Accepted May 24 1994

ISSN:0267-9477    

DOI:10.1039/JA9940901059

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 1063 Analysis of Aluminium Oxide and Silicon Carbide Ceramic Materials by Inductively Coupled Plasma Mass Spectrometry* Invited Lecture J. A. C. Broekaert Universitat Dortmund Fachbereich Chemie 0-44227 Dortmund Germany R. Brandt Max-Planck-lnstitut fur Metallforschung Stuttgart Laboratorium fur Reinststoffanalytik Postfach 722652 0-44073 Dortmund Germany F. Leis C. Pilger and D. Pollmann lnstitut fur Spektrochemie und angewandte Spektroskopie Postfach 707352,D-44073 Germany P. Tschopel and G. Tolgt Max-Planck-lnstitut fur Metallforschung Stuttgart Laboratorium fur Reinststoffanalytik Postfach 122652 0-44073 Dortmund Germany The use of inductively coupled plasma mass spectrometry (ICP-MS) for trace element determinations in A1203 and Sic powders as well as in compact Sic ceramics subsequent to grinding to a particle size of < 20 pm was investigated.The dissolution procedure optimized for A1203 included treatment with HCI and H2S04. For ICP-MS analyses the maximum tolerable AI,O content in the solutions to be measured was found to be 400 pg ml-' for which detection limits in the range 0.002-2 pg g-' were obtained. Analyses of real samples in the concentration range of 0.05 to several hundred pg g-' will be discussed in terms of precision and accuracy and it will be shown that leaching of the powders with acids could provide information on the localization of the impurities. The influence of the removal of CI from the analyte solutions on spectral interferences in the low mass range and of cooling the spray chamber on the power of detection will be discussed.A further method of improving the power of detection could lie in matrix removal based on the on-line complexation of Co Cu Cr Fe Ga Mn Ni and V with hexamethylenedithiocarbamate. For Sic powders dissolution by treatment with HN03 H2S04 and HF will be shown to lead to a number of spectral interferences and to limit the tolerable analyte concentration to 500 pg ml-' for which detection limits range from 0.002 (for heavy elements) to 10 pg g-' for elements such as Mg. Results for the determination of 6 Na Al V Cr Mn Fe Ni Co Cu Ga Sr Y Zr In Sn Ba La Hf Pb and U in ceramic powders of industrial importance will be presented. Keywords Ceramic powders; inductively coupled plasma mass spectrometry; aluminium oxide; silicon carbide Advanced ceramics now find wide interest as working materials in diverse areas of science and technology.Their sometimes extreme properties in terms of hardness corrosion and tem- perature resistance or electrical as well as thermal conductivity often depend greatly on their composition.' In a number of cases this is now known to apply down to the trace element level. Therefore not only is the determination of bulk impurities important but also the micro-distribution of elements in advanced ceramics is crucial. In addition to the characteriz- ation of compact ceramics the analysis of the basic products is also important so as to be able to optimize production procedures including the exclusion of contamination at the various production stages and tailoring the properties of the material by controlling the levels of impurities.For analysis of the basic products atomic spectrometric methods are very important as multi-element characterization down to the sub- pg g-' level is required.2 By applying inductively coupled plasma atomic emission spectrometry (ICP-AES) this goal can be reached for a number of elements in the matrices A1203 Sic and Zr02. Indeed after careful optimization of the working conditions of ICP-AES eventually with the aid of simplex procedures a range of trace elements can be determined in A1203 (Ca Co Cu Fe Mg Mn and Zn),3 Sic (Ca Cd Cr * Presented at the 1994 Winter Conference on Plasma Spectro- t Also at Institut fur Spektrochemie und angewandte Spektroskopie chemistry San Diego CA USA January 10-15 1994.Postfach 101352 D-44013 Dortmund Germany. Cu Mg Mn and Zn)4 and ZrO (Al B Ca Cu Fe Mg Mn Na Ti V and Y),5*6 however severe limitations occur especially for the heavier elements. In the present work it will be shown that for many elements in the case of A1203 and Sic reliable multi-element trace characterization at lower concentration levels can be performed by applying inductively coupled plasma mass spectrometry (ICP-MS) instead of ICP-AES. Herefore the sample prep- aration procedures known from previous ICP-AES work in these laboratories had to be adapted to the requirements of ICP-MS in terms of sample concentrations and also for the acids used during the sample pre-treatment. Special precautions such as cooling of the spray chamber and removal of C1 from the measurement solutions will be discussed for the case of A1203.Also a method of matrix removal which leads to possibilities for further improvement of the power of detection will be presented. In the case of Sic optimization of the dissolution procedure a detailed discussion of spectral inter- ferences in ICP-MS and possibilities for the analysis of compact Sic will be also discussed. Results for commercially available A1203 and Sic materials used for the production of advanced ceramics will be presented. Experiment a1 Instrumentation All measurements were performed with a PlasmaQuad PQ2 Turbo Plus ICP-MS instrument (Fisons Instruments VG1064 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 Elemental Winsford UK) the instrumental data and working parameters for which are listed in Table 1. The analytical parameters used were as proposed for the instrument by the manufacturer. For the direct analyses of the samples subsequent to dissolution pneumatic nebulization with a concentric glass nebulizer (Meinhard Associates Santa Ana CA USA) and also with a Babington-type nebulizer (V-groove nebulizer Fisons Instruments VG Elemental) positioned in a laboratory- made cooled spray chamber7 was used. In the on-line matrix removal using complexation of the trace elements adsorption of the hexamethylenedithiocarbonate (HMDC) complexes on an RP18 column solid-phase extraction and high-pressure nebulization of the effluent were applied. A high-pressure nebulization system (Fa.Knauer Berlin Germany) as devel- oped by Berndt was used in conjunction with desolvation including both water and a Peltier cooling.' A combined knocking-grinding machine with a pestle as well as a mortar made of high-purity Sic (Elektroschmelzwerk Kempten) was used for grinding the compact Sic granulate material. The mortar vessel was held in a steel enclosure (Fig. 1). With this device it was found that Sic granules with grains of between 1 and 20 mm side lengths can be pulverized. Below this grain size the grinding action of the machine was not effective as it seemed that a certain size is required for the pressure to work on the granules. Below this size Sic pieces were found not to be split. Samples and Reagents The A1,0 powders analysed included AKP-20 (mean particle size 0.57 pm as indicated by the manufacturer Sumitomo Japan) and AKP-30 [mean particle size measured by auto- mated electron probe microanalysis (EPMA) 0.35 and 0.43 pm as indicated by the manufacturer Sumitorno] as well as ME/03 (mean particle size measured by automated EPMA 0.35 pm).The ME/03 is a powder which has been characterized previously in a round-robin organized by the Arbeitskreis Refraktarwerkstoffe in the Chemiker Ausschul3 der Gesellschaft Deutscher Metallhiitten- und Bergleute (GDMB). The particle size distribution of the powders was determined by both automated EPMA (for a description of the method see ref. 9) Table 1 Plus ICP-MS instrument Instrumental and working parameters for the PQ2 Turbo ~~ Generator Nebulizers V-groove nebulizer Concentric glass nebulizer Spray chamber Peristaltic pump Interface Sampler Skimmer Power Gas flows Outer Intermediate Aerosol carrier Sample uptake rate Sampling depth Vacuum 1st stage 2nd stage Mass spectrometer Mass range Dwell time Detector Henry 2.0 kW 27.12 MHz Meinhard Associates Scott-type made of quartz Gilson Minipuls 3 Ni aperture 1.0 mm Ni aperture 0.7 mm 1.35 kW forward 10-15 W reflected 14lmin-' 0.9-1.5 1 min-' 0.9- 1.1 1 min - ' (3 bar*) 0.75-1.1 ml min-' 10-12 mm 2.0-3.0 mbar -= 1 x mbar 2.0-3.0 x mbar 10-141 u for A1,0 5-239 u for Sic Dual mode 640 ps pulse counting mode 320 ps Analogue-pulse counting 'L mortar d Rotating vibrator i Steel holder Fig.1 Grinding device for compact Sic and by laser stray radiation measurements (laser particle sizer Analysette 22- Fritsch Idar-Oberstein Germany).Reasonable agreement between the results was obtained considering that deviations possibly can arise from differences in the ultrasonic treatment of the suspension being measured and artifacts from the nuclepore filter loading encountered in the case of EPMA (Fig. 2). The Sic samples analysed were A10 (mean particle size xl pm as indicated by the manufacturer H.C. Starck Goslar Germany) and S-933 (Elektroschmelzwerk Kempten Germany) in the form of a powder (mean particle size ~ 0 . 8 pm as indicated by the manufacturer) and granules (grain size x 1-10 mm). The HCl and the H2S0 used for the sample decomposition were purified by sub-boiling distillation. All dilutions were 20 I 100 80 I 60 40 20 0 (b) 0.1 oF.o 2.0 / 20 A 10 B 0 0.5 1 5 10 D i a m ete r/pm Fig. 2 Particle size distribution obtained for the A1,0 powder AKP-30 (Sumitomo Japan) by (a) EPMA (mean diameter = 0.35 pm); and (b) laser light scattering (mean diameter=0.59 pm) * 1 bar= 1 x lo5 Pa.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 made with H20 doubly distilled in quartz equipment. The HF used (Merck Darmstadt Germany) was of Suprapur quality. For the preparation of the standard solutions the respective Titrisol stock solutions (Merck) were used. Sample decompo- sitions were performed at high temperature and pressure in closed poly( tetrafluoroethylene) (PTFE) vessels (Berghof DAB 111 Tubingen Germany). In the matrix removal studies HMDC (Merck) and C RP18 columns (Fa.Knauer) were used. Results and Discussion Analysis of A1203 Powders After initial tests with an AlCl solution it was found that the analyte solution for ICP-MS investigations should not contain more than ~ 2 0 0 pg ml of Al which corresponds to a concen- tration of 400 pg ml-' of A1,0,. Indeed at higher concen- trations high ablation of the sampler was observed and also a glassy deposit was soon formed on the skimmer. Both lead to clogging as well as to a deterioration of the short-term precision expressed by the relative standard deviations (RSDs) and long-term drifts. 15 A 1 3t 0 50 100 150 200 250 Time/mi n Fig.3 Long-term stability in ICP-MS analyses of A1203 after acid decomposition A 24Mg; B 66Zn; C '39La; D,139 La:'I5In and E 24Mg 45Sc. Sample solutions 10 ng ml-' of B Na Mg Ti V Cr Mn Co Ni Zn Ga Zr Ba La and Ce; 400 pg ml- of A1203 10 ng ml-' In and 50 ng ml-' of Sc in 0.04 moll- HC1-0.0064 moll- ' H2S04. ICP-MS PQ2 Turbo Plus with quartz spray-chamber (10 "C) (measurement conditions as in Table 1 ).Measurement cycle pre-flow 90 s; measurement time 5 x 60 s; rinsing with acid solution 2.5 min 1065 At a sample concentration of 400pgml-' of A120 the RSDs were 2-5% for impurity concentrations of 10 ng ml-' and in a 0.04mo11-1 HC1 and 6 . 4 ~ lop3 moll-' H2S04 solution which are the concentrations of acid present after acid decomposition of the samples. However when adding 10 ng ml-' of In and 50 ng ml-' of Sc as internal standards the short-term precision for all elements investigated (B Na Mg Ti V Cr Mn Co Ni Zn Ga Zr Ba La and Ce) becomes of the order of 1-2% and drifts [with an intermediate washing stage of 2.5 min after each sample (measurement cycle pre-flow 90 s integration 5 x 60 s)] after an initial period are below 5% over a period of 4 h (Fig.3). For dissolution of the Al,O powders treatment with HC1 and H,SO at 225°C according to the following procedure was found to be optimum. A 1 +0.001 g sample of A120 powder was transferred into a 150 ml PTFE vessel (Fa. Berghof) and 10 ml of sub-boiled HCl plus 1 ml of sub-boiled H,SO and 5ml of H20 were also added. The mixture was allowed to react for 6 h at 225 "C in the closed vessel. The resulting solution had been diluted 1 + 2500. Four solutions with matched acid concentrations and analyte concentrations of 0.4 4 40 and 400 ng ml-' and containing 5 ng ml-l of Rh as an internal standard were used for calibration purposes.The rinsing solution used contained 4ml of sub-boiled HCl and 0.4 ml of sub-boiled H,S04 in 1 1 of H,O. For analyte concentrations of 400 pg ml-' of A1203 the detection limits obtained for ICP-MS (based on the 3s cri- terion) with a non-cooled spray chamber and considering the blank limitations of all reagents used are listed in Table2. They range from 0.1 to 4 pg g-' and are higher for the low mass elements. However severe blanks (Na) or spectral inter- ferences (Ca Fe and Zn) are seen to occur. In the case of the instrument used no systematic differences could be noticed when comparing the detection limits obtained in the scanning mode with those of the peak jumping mode for a dwell time of 640ps per channel and 25 scan cycles.Furthermore the increased values found for a number of elements were clearly due to spectral interferences by oxygen chlorine or sulfur containing cluster ions of argon (see Table 3). Under the conditions originally used the detection limits with ICP-MS for a number of elements with respect to the solid samples were even lower than with ICP-AES however only for elements with a mass below 60. At higher masses this situation changes as with ICP-AES it is no longer possible to determine elements of interest in the concentration range required for the analysis of advanced ceramics. Table 2 Detection limits (cL) (3s concept) of ICP-MS and ICP-AES for the analysis of A1203. Values in parentheses are for analyte solutions in distilled H20.Other values based on four replicates (ICP-MS) or 12 replicates (ICP-AES) of the decomposition with all acids ICP-MS * Scanning mode c,/ng ml - ' 1.0 (0.25) 6.5 (0.4) 2.0 (1.9) 32 (6.4) 0.8 (0.4) 0.3 (0.1) 36 (4.3) 0.3 (0.07) 2.0 (0.4) 0.6 (1.4) 0.06 (0.03) 0.05 (0.04) - (0.01) - (0.02) 0.01 (0.01) - (0.01) Peak jumping c,/ng ml-' 1.3 (0.1) 8.3 (1.3) 1.4 (1.4) 30 (3.8) 1.4 (0.1) 0.3 (0.06) 35 (2.6) 0.08 (0.06) 0.5 (0.07) 0.07 (0.07) 1.8 (0.2) 0.8 (1.1) 0.13 (0.07) 0.16 (0.16) 0.06 (0.05) 0.07 (0.05) CLIPg g - 3.2 18.0 3.5 75 3.5 0.75 0.75 1.2 0.18 4.5 1.5 0.32 0.32 0.15 0.17 87 Actual? c,,lClg g - 0.6 1.1 0.7 1.2 0.04 0.024 0.08 0.04 0.8 0.008 0.23 0.002 0.008 0.008 ICP-AES CJPg !3 - 0.6 0.03 0.05 0.3 0.8 2.1 0.9 ~~~~~~~ ~ ~ * Sample dilution 1 +2500; PQ2 Turbo Plus.7 Values obtained after cooling the quartz spray chamber down to 10 "C. $ From ref. 3. Sample dilution 1 + 50; 2 kW ICP 0.9 m Czerny-Turner monochromator.1066 JOURNAL OF ANALYT1CA.L ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 Table 3 with HCl-H2S04 in a PTFE vessel Spectral interferences in ICP-MS for Alz03. Decomposition Analyte 44Ca 47Ti 48Ti 51v "Cr 53Cr "Mn 56Fe "Fe 60Ni 64Zn 66Zn 71Ga Interferen t 12~160160 2 7 ~ 1 1 6 0 ~ 3 3 ~ 1 4 ~ 3 2 ~ 1 6 0 3 5 CPO 3 7 ~ 1 1 6 0 4 0 ~ ~ 1 6 0 35C1160H 40Ar12C 36Ar160 36S160 40Ar'4NH 40Ar160H Cones 3 2 ~ 1 6 0 1 6 0 34~160160 3 2 ~ 3 4 s 36Ar35C1 In the case of the AKP-30 powder results in the 0.1-2 pg g-' range were obtained for B Ti Cr Mn Ni Cu Zr Ba La and Ce (Table 4).The results obtained in the peak jumping mode agreed well with those of the scanning mode. Even with the semiquantitative programe the results deviated by less than a factor of two with only a few exceptions. In the case of Fe and Ca however severe deviations were obtained which could relate more to interference problems as these are known already from the early literature on ICP-MS (see for example ref. lo) than to blank limitations. The accuracy of the ICP-MS results obtained was evaluated by performing analyses of the ME/03 powder which has also been analysed extensively by other techniques (Table 5). Excepted for Fe reasonable agree- Table 4 Results for analysis of Al,03 powder (AKP-30) by ICP-MS. All concentrations are in pg g-' f the standard deviations from four replicate analyses; final dilution 1 + 2500; and internal standard Rh Anal yte llB 47Ti "Cr "Mn 60Ni 63cu 71Ga 'OZr 138Ba I3'La l4OCe Semiquantitative Scanning mode Peak jumping 2.2 f 0.2 4.5 10.6 4.8 f0.3 1.3f0.3 0.15 f O .l - 2.6 _+ 0.2 1.3f0.15 1.0 f 0.07 1.1 fO.1 0.7 f 0.15 0.25 f 0.09 1.5f0.1 0.7 f 0.05 1.1 f0.04 - 0.05 f 0.02 < 0.08 0.4 f 0.2 0.4 f 0.25 0.5 f 0.3 1.0f0.3 1.1 k0.2 1.1 k0.3 0.14f0.01 0.16f0.01 0.13f0.01 0.4 f 0.02 0.5 f 0.02 0.4 f 0.02 0.12 f 0.06 - - ment of the ICP-MS results with those of other methods was obtained at levels ranging from concentrations of sub-pg g-' to 700 pg g-' which clearly illustrates the analytical features Further possibilities for ICP-MS are shown by the results from leaching experiments performed with the AKP-20 and AKP-30 powders.Indeed when leaching the samples with dilute HNO (22 ml of sub-boiled HNO in 11 of H20) overnight and then analysing the leaching liquid and the residue after acid decomposition by ICP-MS the sums of the results obtained are in good agreement with those from independent methods (Table 6). This shows that ICP-MS should be of great use for monitoring sources of contamination during the stages of powder processing but also that it could be of use for showing differences in the localization of the different elements in the individual grains which could again indicate their origin. Based on experience of the limitations of the actual power of detection several measures could be taken so as to realize progress in detection limits for Alz03.The first improvements in the power of detection resulted from the use of a cooled spray chamber. This measure was shown to lead to a reduction of the white noise in particular as described el~ewhere.~ This leads directly to an improvement in the short-term fluctuations as expressed by the RSDs but also to a reduction of the spectral interferences as reflected by the signals obtained on the different channels for doubly distilled HzO (Table 7). Accordingly it was found that the power of detection could be considerably improved by cooling the spray chamber down to 10°C. The use of a quartz spray chamber and rinsing of the spray chamber injector tube assembly by steaming with vapours of HNO brought further improvement. The improved detection limits for the A1203 samples taking into account the blanks obtained with a cooled and pre-cleaned quartz spray chamber in the pulse counting mode are given in a separate column of Table 2.It was also found that for a number of powders the use of H,SO was not strictly required. Indeed with the same reaction time (6 h) it was found that the decomposition could be performed with 10ml of sub-boiled HC1 and 5ml of HzO added to 1 g of sample provided the temperature was increased from 225 to 240"C as experienced with the AKP-20 and AKP-30 powders. It was further found that the interferences could be considerably reduced when Cl was removed from the sample solutions by treatment with sub-boiling HNO,. When working under clean-room conditions recoveries for Rb Sr Y Zr Rh La Ce Sm Eu W and Th at the 20ngml-' level of ICP-MS.Table 5 analyses. Final dilution 1 + 2500; internal standard Rh Results for analysis of A1203 powder ME/03 by ICP-MS. All concentrations are in pg g fthe standard deviations from four replicate ICP-MS Comparison with ICP-AES Anal yte llB 23Na 24Mg 44Ca 52Cr "Mn 57Fe 59c0 60Ni 63cu 66Zn 71Ga 'OZr 138Ba 13'La "OCe Scanning 0.4 f 0.2 781 f 33 492 f 21 404 k 30 1.5fO.l 1.6f0.02 359 f 20 0.4 f 0.02 10.2 f0.5 1.4 f 0.1 3.8 k0.5 87 f 2.4 7.5k0.15 20.2 & 2.8 0.03 & 0.01 0.1 Peak jumping 735 f 32 464 f 20 268 38 1.0 & 0.4 1.2k0.2 326 f 8 0.4 f 0.01 9.7 f0.3 1.3f0.05 3.0 f 0.08 90f 1.3 6.6 f 0.6 17.3 f 2.0 0.01 f 0.01 0.07 f 0.02 - Our work <3 777 456 f 7 191 f 10 1.4 k 0.01 251 k 3 <6 9.0 f 0.2 <3 < 13 92 f 0.8 - - - - - Other laboratories 2.2 k 0.5 755 +42 463 k 16 178 k 17 3.3 f 1.9 2.0 f 0.7 245 k 10 11.4 f 5.9 1.0 * 0.1 6.2f4 87+5 9.2f 1.0 1.1 - 0.1* 0.3* * Results from neutron activation analysis.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 1067 Table 6 Determination of the leachable and the non-leachable impurities in the A1,0 samples AKP-20 and AKP-30 by ICP-MS when leaching with 2% (v/v) HN03. Figures in parentheses are the standard deviations from four replicate analyses. The results given for the alternative methods are from acid decomposition of the sample without applying leaching with 2% HNO (v/v) Concentration/pg g- ' Analyte B Na Mg Ca Ti V Cr Mn Fe c o Ni c u Zn Ga Zr Ba La Ce AKP-20- AKP-30- B Na Mg Ca Ti V Cr Mn Fe c o Ni c u Zn Ga Zr Ba La Ce Leachable 2.8 (0.05) 2.8 (0.09) 0.98 (0.007) 0.07 (0.001) 0.2 (0.003) 0.15 (0.004) 0.026 (0.0003) 1.0 (0.1) 0.002 (0.00005) 0.1 (0.002) 0.55 (0.005) 0.33 (0.01) 0.006 (0.0003) 0.015 (0.0003) 0.046 (0.003) 0.037 (0.0005) 1.1 (0.02) 0.1 (0.002) 0.82 (0.02) 1.8 (0.04) 1.2 (0.3) 1.02 (0.02) 0.1 (0.001) 0.1 (0.002) 0.09 (0.002) 0.017 (0.0005) 0.8 (0.1) 0.002 (0.0003) 0.07 (0.001) 0.62 (0.003) 0.33 (0.002) 0.01 (0.0005) 0.01 5 (0.0003) 0.22 (0.004) 0.29 (0.005) 0.11 (0.002) Non-leachable 6.5 < 1.1 1.1 0.33 1 .o 1.9 0.16 9.9 0.04 0.31 0.57 < 0.8 0.17 < 0.23 0.01 0.04 0.17 < 24 2.3 < 1.1 1.6 0.19 1.5 1.4 0.10 6.8 0.11 0.21 1.0 < 0.8 0.08 < 0.23 0.02 0.17 0.52 < 24 Sum 9.3 < 3.9 2.2 1.31 0.84 2.05 0.19 0.042 0.41 1.12 < 1.13 0.18 < 0.245 0.056 0.077 0.27 < 24.2 10.9 3.12 < 2.9 2.8 1.21 1.6 < 24.1 1.49 0.117 7.6 0.1 12 0.28 1.62 < 1.13 0.09 < 0.245 0.24 0.28 0.8 1 Alternative method 8.2 (ICP-AES) 4.0 (NAA)* 2.4 (ICP-AES) 1.26 (ICP-AES) 0.77 (ICP-AES) < 10 (TXRF)? 1.35 (NAA) < 0.3 (ICP-AES) 10.2 (ICP-AES) 0.032 (NAA) <3 (TXRF) 1.6 (NAA) 0.12 (NAA) 1.1 (NAA) 0.087 (NAA) 0.24 (NAA) 1.4 (ICP-AES) - 3.3 (ICP-AES) - 2.06 (ICP-AES) 1.26 (ICP-AES) 1.06 (ICP-AES) - 1.0 (ICP-MS) < 0.3 (ICP-AES) 7.1 (ICP-AES) < 2.1 (ICP-AES) 0.25 (ICP-MS) 1.29 (ICP-AES) - < 0.08 (ICP-MS) 0.5 (ICP-MS) 1.1 (ICP-MS) 0.1 3 (ICP-MS) 0.4 (ICP-MS) * NAA neutron activation analysis.TXRF total-reflection X-ray fluorescence. and in the presence of 400 pg ml-I of A1,0 were better than loo+ 10%.However for a number of elements the losses soon became larger with this procedure (Table 8). Another method of improving the power of detection could result from separation of the trace elements to be determined from the aluminium matrix as is possible by on-line separation from the complexes of a number of elements with dithiocarb- oxylates such as HMDC. This can be realized with the aid of a high-performance liquid chromatography (HPLC) system where the HMDC complexes are sorbed onto a suitable column and released by solid-phase extraction. The eluate can then be brought into aerosol form by high-pressure nebuliz- ation desolvated and released into the ICP-MS instrument." In the present work complexation of the trace elements was performed by adding 500 pl of a 4% HMDC solution to the decomposition solution.Subsequently dilution to a final volume of 100 ml was performed and the pH was adjusted to 2-3 with concentrated ammonia solution. When passing 1 ml of the solution through a C18 RP column and using a methanol-H,O mixture of 70 + 30 as the eluent recoveries of better than 97% could be found for a number of trace elements (V Fe Ni Co Cu and Ga) at the 100ng ml-1 level. Over 99.5 +0.3% of the A1 matrix could be removed at the 2 mg ml-1 concentration level. These measures certainly will enable a considerable improve- ment in the actual power of detection to be realized for the case of Al,O,. With the present state-of-the-art improvements in the power of detection of ICP-MS with respect to the one achieved with ICP-AES for the low mass range are possible whereas up till now for some elements the actual power of detection of ICP-MS obtained with respect to solid samples was not improved as compared with ICP-AES (Table 2).Such progress is clearly required in view of the requirements for the determination of a number of elements in powders used for the production of advanced ceramics. Analysis of Sic Powders For the case of Sic powders sample decomposition prior to ICP-MS analysis was found to be achieved efficiently by treatment of the powders with a mixture of HF HNO and H2S04 according to the procedure described below. For the decomposition of Sic 250mg of powder were treated for 12 h at 250°C with 4 ml of concentrated HF (Suprapur) and 4 ml of concentrated HNO as well as 4 ml of concentrated H2S04 both of which were purified by sub- boiling distillation.The resulting solution was then diluted to 500 ml and accordingly the solution used for measurement contained 500pgml-' of Sic in 0.8% HNO 1.5% H2S04 and 0.5% HF. Analyses by ICP-MS included five measure-1068 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 Table7 aqueous standard solution 20 ng ml-'; and short-term stability (30 min) Influence of a cooled spray chamber made of quartz on the stability and on the blank signal levels in ICP-MS. Meinhard nebulizer; Without cooling With cooling to 10°C Standard deviation 3.0 (Yo) - - 2.2 - 2.2 2.0 - - 2.5 - 3.1 - 3.1 3.5 3.7 3.4 Blank concentration*/ ng ml-' - 0.7 22.0 2.1 0.11 0.03 0.56 0.03 0.75 0.37 0.44 0.1 1 0.01 0.03 0.02 0.02 112 - - Standard deviation 2.5 W O ) - 2.7 - 2.3 1.8 - 2.3 - 1.3 - 1.3 1.4 1.6 2.0 Blank concentration/ ng ml-' 0.5 3.00 0.7 0.07 0.02 0.20 0.004 0.11 0.13 0.26 0.1 1 0.04 0.12 0.005 0.008 - 14 - - * Blank signal for doubly distilled H,O converted into concentrations.Table 8 Recoveries obtained for a series of trace elements on removal of C1 from AlCl,-containing solutions Procedure involved dissolution of 3.3 g of A1Cl3.6H2O in 18 ml of H20 and addition of 1 pg of the trace elements addition of 2 ml sub-boiled HN03 adjusting to 50 ml with H20 and evaporation of the sample on a heating plate under clean-room conditions Element Be Ti Mn Co Rb Sr Y Zr Mo Rh La Ce Sm Eu W T1 Mg Concentration found/ ng ml-I 15.72 17.84 16.45 17.81 17.49 18.45 19.00 19.12 18.85 17.86 18.15 18.69 19.25 19.98 19.29 20.85 19.71 Recovery W) 78.6 89.2 82.3 89.1 87.5 92.3 95.0 95.6 94.3 89.3 90.8 93.5 96.3 99.9 96.5 104.3 98.6 ments for each sample and calibration was achieved with four independent blanks and five standard solutions.In order to carry this out the ICP Multielement Standard IV (Merck) was used to which Ti V Zr Zn La Hf and U were added and of which dilutions of 0.1 0.5 1 10 and 100 ng ml-1 were made with the addition of 40ngml-' of Rh as an internal standard. A sample concentration of 500 pg ml-1 appeared to be the maximum tolerable matrix concentration because of the risk of deposition on the cones and especially on the skimmer. The analytical short-term precision was found to be between 1 and 3% depending on the mass and concentration.At the maximum sample concentration of 500 pg ml-' the detection limits found for Li B Na Mg Al V Cr Mn Fe Ni Co Cu Zn Ga Sr Y Zr Ag Cd In Sn Ba La Hf T1 Pb Bi and U were found to be between 0.002 and 17 pg g-' (Table 9). The high values for Fe Zn and Mg certainly relate to spectral interferences whereas for Na contamination could be the predominant problem (Table 10). It could be shown however that for the case of the Sic powder S-933 the ICP-MS results obtained subsequent to dissolution agreed well with results for ICP-AES after sample dissolution as well as with those from instrumental NAA ( INAA),12 the ICP-AES suspension tech- niqueI3 and electrothermal atomic absorption spectrometry (ETAAS) using slurry atomi~ation,'~ for all of which the detection limits are listed in the same table.In the case of ICP-MS for Sic powders it could be expected that removal of the Si and the C arising from the matrix as well as of the decomposition acids by heating the solutions to near dryness could improve the limit of detection. However whether such procedures would lead to high analyte losses for a number of relevant elements must be investigated. Further progress in the analysis of Sic powders by ICP-MS could result from the application of slurry nebulization which however could be hampered by abrasion of the sampler and deposits on the skimmer hence also limiting the sample concentrations in the slurries. The application of electrothermal evaporation as applied in ICP-AES for ceramic powders by Nickel et could be useful as well but this would certainly necessitate detailed investigations of the calibration.In order to make progress in the analysis of compact ceramics the possibilities of advanced grinding devices could be useful. With a device made of high-purity Sic (Fig. l) it has been found that it is possible to grind a Sic granulate material (1-20 mm side lengths of the grains) in less than 15 s to a powder with a particle size of <20 pm for the majority of the milled sample. However possible contaminations during the grinding procedure will have to be investigated so as to evaluate the feasibility of the device as compared with other direct solids procedures such as laser ablation ICP-MS. Conclusion The results presented for A120 and Sic powders show that ICP-MS is a very powerful tool for multi-element trace charac-JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 1069 Table 9 Detection limits for Sic and results for analysis (pg g-') of Sic powder S-933 obtained with ICP-MS (present work) INAA and ICP- AES after sample decomposition and ETAAS using slurry atomization. Figures in parenthesis are the cL (pgg-') based on the 3s concept including blank limitations. Sample S-933. Values & are the standard deviations calculated from four replicate analyses and five repetitive measurments of each sample solution obtained by acid decomposition ICP-MS decomposition < CL (0.034) 2.7 f 0.5 (0-9) 35.8 f 1.2 (7.3) < CL (9.5) (6) Interference 192f 15 Interference 80.7 f 2.0 (0.5) 8.2 kO.8 (0.4) 7.8 f 0.6 (0.4) 0.79 f 0.05 (0.06) 348f13 (17) 5.6 f 0.4 (0.2) 0.079 f 0.01 3 (0.01) 3.37 f 0.24 (0.09) < CL 0.082 f 0.01 1 (0.03) 0.061 f 0.01 5 (0.023) 0.066 f 0.003 (0.006) 18.5 f0.4 (0.29) < CL < CL 0.0088 k 0.0024 (0.0027) 2.8 f 0.4 (0.06) 3.0 f 0.3 (0.04) 0.23 & 0.03 (0.11) 0.22 f 0.03 (0.08) 0.24 f 0.03 (0.05) 0.14 f 0.02 (0.005) 0.99 f 0.03 (0.02) 1 .OO f 0.05 (0.06) < CL (0.005) (3.3) (0.6) (0.02) ICP-OES Decomposition* < CL (0.15) 2.7 f 0.2 (0.6) 36.1 f 0.4 (2-5) 3.7 f 0.2 (0.02) 5.5 f 1.34 178&4 (2.5) 1702 13§ 9.71 f 0.34 (2.3) 9.6 f 0.q 158f 1 (0.6) 131 f 59 83.7 f 1.8 (2) 6.9 f 0.5 (1.5) - 0.28 f 0.01 (0.06) INAA* - - 35.9 f 0.1 (0.05) - - < CL1 (60) - - 7.1 & 0.2 (0.01) - 0.70 f 0.01 (0.004) 322 & 14 (0.6) 3.8 k0.2 (0.1) 0.063 f 0.002 (0.0005) <CL ( 5 ) < CL (0.5) < CL (0.7) < CL (0.29) < CL (0.5) 21.4f 1.1 (1) 0.021 f 0.003 (0.002) < CL (0.1) < CL (0.03) 3.6 k 0.4 (0.4) - < CL (1) - - 0.1 7 f 0.02 (0.0005) 0.58 f0.03 (0.0008) - - ETAAS slurry atomization' - - - 5.2 f 0.8 177& 19 - 130f 13 ( 5 ) (2) (0.2) 84f 1 7.1 k0.5 0.72 f 0.08 (0.02) 340 L 20 (2) 4.8 k 0.7 (0.8) 3.2 k 0.6 (0.05) 0.3 1 & 0.03 (0.01) - - - - - < CL (0.02) - - - - - - - - - - (continued on next page)1070 Table 9 (continued) JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 206Pb 2.36 k 0.06 ’07Pb 2.54 Ifi 0.12 ’08Pb 2.51 kO.11 (0.04) (0.05) (0.04) < CL (0.004) (0.002) 209~i 2 3 8 ~ 0.043 & 0.002 < CL (1.6) 0.085f0.017 (0.005) * Ref.12. t Ref. 13. $ Ref. 15. fj Measurements from our laboratories using ICP-AES. Table 10 Spectral interferences in ICP-MS for Sic. Decomposition with H2S04-HN03-HF in PTFE vesselsL4 terization of ceramic powders. However restrictions arise from the low sample concentrations that can be tolerated. Because of this and also of the spectral interferences serious limitations of the actual levels of detection were found especially for the light elements which are very important for advanced ceramics. These limitations however certainly in part can be overcome with appropriate measures as has been indicated and are under further investigation. The work presented has been supported by the ‘Deutsche Forschungsgemeinschaft’ (Bonn) within the framework of the ‘SPP Keramische Hochleistungswerkstoffe’. 1 2 3 7 8 9 10 11 12 13 14 15 16 References Broekaert J.A. C. Graule T. Jenett H. Tolg G. and Tschopel P. Fresenius’ 2. Anal. Chem. 1989 332 825. Broekaert J. A. C. and Tolg G. Mikrochim. Acta 1990 11 173. Graule T. von Bohlen A. Broekaert J. A. C. Grallath E. Klockenkamper R. Tschopel P. and Tolg G. Fresenius’ Z. Anal. Chem. 1989 335 637. Docekal B. Broekaert J. A. C. Graule T. Tschopel P. and Tolg G. Fresenius’ J. Anal. Chem. 1992 342 113. Lobinski R. Broekaert J. A. C. Tschopel P. and Tolg G. Fresenius’ J. Anal. Chem. 1992 342 569. Lobinski R. Van Born W. Broekaert J. A. C. Tschopel P. and Tolg G. Fresenius’ J. Anal. Chem. 1992 342 563. Pollmann D. Pilger C. Hergenroder R. Leis F. Tschopel P. and Broekaert J. A. C. Spectrochim. Acta Part B in the press. Jakubowski N. Feldmann I. Stuwer D. and Berndt H. Spectrochim. Acta Part B 1992 47 119. Raeymaekers B. J. Van Espen P. and Adams F. Mikrochim. Acta 1984 11 437. Tan S. H. and Horlick G. Appl. Spectrosc. 1986 40 445. Pollmann D. Leis F. Tschopel P. Broekaert J. A. C. and Tolg G. poster ThP41 presented at the 1994 Winter Conference on Plasma Spectrochemistry San Diego CA USA January Franek M. and Krivan V. Fresenius’ J. Anal. Chem. 1992 342 118. Lathen C. Ph.D. Dissertation University of Dortmund 1992. Broekaert J. A. C. Brandt R. Lathen C. Pilger C. Pollmann D. Tschopel P. and Tolg G. Fresenius’ J. Anal. Chem. 1994,349,20. Docekal B. and Krivan V. J. Anal. At. Spectrom. 1992 7 521. Nickel H. Zadgorska Z. and Wolff G. Spectrochim. Acta Part B 1993 48 25. 10-15 1994. Paper 4/00375F Received January 20 1994 Accepted March 8 1994

ISSN:0267-9477    

DOI:10.1039/JA9940901063

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

JOURNAL O F ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 107 1 Comparative Study of Multi-element Determination Using Inductively Coupled Plasma Mass Spectrometry Total Reflection X-ray Fluorescence Spectrometry and Neutron Activation Analysis* Rudolf Pepelnik Andreas Prange and Rainer NiedergesaO hstitute of Physics GKSS Research Centre Geesthacht GmbH P.O. B. 7 760 0-27494 Geesthacht Germany The analytical capabilities of inductively coupled plasma mass spectrometry total reflection X-ray fluor- escence spectrometry and neutron activation analysis were compared. The data originated from the parallel analyses of spinach cabbage and domestic sludge samples which were used in inter-laboratory tests to monitor precision and accuracy. The determined concentrations range from 40 mg 9 - l to 20 ng 9 - l and the analytical errors from 2 to 30%. The extensive results and the reliability of the techniques are discussed.Keywords Inductively coupled plasma mass spectrometry; total reflection X-ray spectrometry; neutron activation analysis ; m ulti- elem en t analysis ; biological samples Inductively coupled plasma mass spectrometry (ICP-MS) total reflection X-ray fluorescence spectrometry (TXRF) and neu- tron activation analysis (NAA) were compared for their suit- ability for application to the multi-element analysis of spinach cabbage and domestic sludge samples. Analytical Techniques The most important characteristics of the analytical techniques used are summarized in Table 1. There are differences in sample preparation calibration and analysis times.For ICP-MS a Perkin-Elmer Sciex Elan 5000 instrument with a quadrupole mass separator and a channel electron multiplier was applied. The argon plasma was operated at 1000 W. The digested samples were introduced by a peristaltic pump and a cross-flow nebulizer at a rate of about 1.2 ml min-'. An ICP multi-element standard solution (Merck IV) containing 23 elements a rare earth standard solution with 16 elements and several standards of single elements were used for calibration. Each element has to be calibrated using adapted standard solutions according to acid- ity and concentration in order to obtain accurate results. As internal standard a solution containing Rh was used during the entire analysis. TXRF is a special variant of the energy-dispersive XRF method.An Extra I1 device from Atomika Instruments (Oberschleissheim Germany) with an Si(Li) detector and a multi-channel analyser were used Molybdenum and tungsten X-ray tubes were operated at 38 mA and 49 kV. A drop of the digested sample typically 10 pl was placed on a quartz glass sample carrier dried and excited by totally reflected X-rays.'q2 Owing to the total reflection the sample was excited twice and almost no background was produced below the X-ray fluor- escence lines. The method can also be used for very small sample volumes. NAA is in principle a non-destructive analytical technique. No pre-treatment of samples is necessary. Consequently the completeness of digestion and any contamination or loss during sample preparation such as occasionally occurs in other methods can be tested by comparison with this technique.Dry aliquots of the samples (90-200 mg) were irradiated in the GKSS FRG-1 research reactor at a thermal neutron flux of 2 x 1013-5 x 1013 cmU2 s-I. As neutron flux monitors Fe Ni and Au foils were irradiated with the sample^.^ The irradiation times varied from 5 min to 3 d the decay times from 6 min to * Presented at the 1994 Winter Conference on Plasma Spectro- chemistry San Diego CA USA January 10-15 1994. 30 d and the counting periods from 5 min to 12 h. For pray spectrometry an HPGe detector a multi-channel analyser and a computer program4 were used. Mg Al C1 Ti V Mn and Cu were determined after a short irradiation and the other elements by long-term activation. Comparison of the Techniques ICP-MS is a very sensitive and rapid method.It is applicable to the detection of almost all elements. The detection limits are in the range 1-100 ng 1-1 for most elements. The accuracy of the determination of Si P S K Ca Se and Br is adversely affected by blank values due to combinations of argon oxygen nitrogen and hydrogen. For some elements the probability of forming positive ions is low (e.g. the degree of ionization for fluorine is 0.0002). The disadvantages of ICP-MS are the comprehensive calibration procedures involved which are necessary for all elements the spectral interferences caused by the main constituents of the gases and other matrix elements (e.g. As) and the limitation to low salt contents (below 0.5%). Highly concentrated solutions have to be diluted.For several elements such as mercury and boron memory effects in the sample introduction system limit the accuracy in routine analysis. Up to 45 elemental concentrations have been deter- mined by ICP-MS. TXRF is a powerful multi-element technique using a very simple calibration for all elements which is independent of the sample matrix. The sample is mixed with a known amount of internal standard (e.g. a solution of cobalt). One internal standard of a single element is sufficient to achieve absolute concentration values for all elements and all matrices using a computer program. There is no limitation due to memory effects or acid concentration because for each sample a separ- ate sample carrier is used. The detection limit is of the order of 1Opg for 70 elements but it depends on the mass of the matrix.Thus the detection power of this method is high for the analysis of water samples but it is limited in the analysis of ultra-trace amounts of heavy solids that is if no initial chemical separation is performed. One disadvantage of XRF in general is that the detection of light elements with an order number below 13 is hindered by the high absorption effects of the low X-ray energies of these elements. Another limitation is the determination of neighbouring elements that have concen- tration differences of more than three orders of magnitude owing to the overlapping X-ray lines. Up to 38 elements have been determined by TXRF. The analytical sensitivity of NAA is very high for most elements. It depends on the nuclear data and reaches the sub- picogram level.5 There are almost no spectral interferences1072 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 caused by the matrix apart from a possibly high background of an element with a large activation cross-section (e.g. Na). Some elements e.g. B P and S and Pb are not measurable without additional devices or chemical procedures. Lead can be determined by Pb(n,~)"~"Pb using fast neutrons and a fast rabbit system. The method needs an intense neutron source (e.g. a nuclear reactor) and is time consuming because of the relatively long irradiation and decay times. Instrumental NAA involves less operator time than other techniques. Up to 45 elements have been determined by NAA. Table 1 Characteristics of the analytical techniques Characteristic Amount of sample Sample preparation Typical dilution Sample volume Sample mass Instrument Excitation Calibration Analysis time Total time Limitations Advantages Disadvantages ICP-MS 200-300 mg Pressure digestion sub-boiled HN03 1 1000 2-5 ml Perkin-Elmer Sciex Elan 5000 Ar plasma lo00 W - External multi-element standard 2-3 min internal standard 8 h C F S C1 Ca Fast analysis High sensitivity Li-U Calibration for each element Spectral overlaps by Ar and matrix Memory effects TRXF 200-300 mg Pressure digestion sub-boiled HN03 1 100 10-20 pl - Atomika Instruments Extra I1 Mo W X-rays 38 mA 49 kV Internal single-element standard 20 min 8 h Light elements Z < 13 Small sample volumes Simple calibration High sensitivity No interference by matrix No memory effects matrix Influenced by neighbouring elemem ts Detection limits depend on mass of NAA 100-200 mg No pre-treatment None 20-200 mg FRGl research reactor Thermal neutron capture neutron flux (2-5) x 1013 n cm-2 s-' External neutron flux monitors - Irradiation 5 min-3 d Decay 5 min-30 d Measurement 5 min-9 h 15 min-33 d B C P S Pb Non-destructive No losses by digestion High sensitivity No memory effects Nuclear reactor Time consuming Table 2 Elemental concentrations in NIST SRM 1570 Spinach (mg kg- ').Results are means +standard deviations (n=6) Element B Na Mg A1 P S c1 K Ca s c Ti V Cr Mn Fe c o Ni c u Zn As Br Rb Sr Mo Cd Sb cs Ba La Ce Nd Sm Hf W Hg Pb Th U 870 & 50 5500 f 200 - 35600 k 300 13500f 300 (0.16) - 4.6 f 0.3 165k6 550 f 20 (1.5) (6) 12$-2 50&2 0.15f0.05 (54) 12.1 10.2 87f2 (1.5) - (0.04) - - (0.37) - - 0.030 -t 0.005 1.2f0.2 0.12&0.03 0.046 f 0.009 Consensus 27.7 f0.6 14200_+1000 8650 f 3 10 810+90 5240f310 4350f470 6500 + 410 35600f 1500 13300f800 0.17f0.01 18f10 1.20k0.16 4.3 k 0.5 164+6 540 f 30 1.56 f 0.12 5.6 & 0.7 11.8f0.7 50f4 0.15f0.02 48+4 1 1.5 f 0.9 80f5 0.30 & 0.08 1.43 k0.14 0.040 f 0.009 0.061 f 0.009 14.9 f 2.5 0.34 & 0.04 0.46 (0.31) 0.056 f0.024 (0.14) 0.030 & 0.004 1.19 & 0.25 0.12 k 0.03 0.046 + 0.003 (0.04) ICP-MS 31f1 140OOf 1000 8600 + 400 700 f 160 4000 k 300 38OOOf 1600 13300 & 500 0.21 k 0.06 29f3 1.2f0.1 4.8 f 0.5 165f4 540 4 16 6.0 f 0.4 11.1 f0.5 51k3 0.1510.02 10.8 1 1.2 8712 0.29 f 0.04 1.3f0.2 - - - - - - 13.24 1.3 0.28 f 0.08 0.41 k0.10 0.11 f0.04 l.OfO.1 0.11 f0.03 TXRF 5000 + 400 4010 & 240 35500 + 1400 13 100 f 750 - - 40f 12 6.6 0.6 164+4 544 & 14 6.3 _+ 0.3 12.0 f 0.4 52f3 - - - 12.3 +0.4 8713 1.610.1 - - - 13.3 & 1.6 - - 1.1 20.2 - NAA - 14500 +_ 360 8360 f 210 865 f 61 - - 7350+ 180 35800 & 900 11900 & 300 0.15 f 0.01 1.26 f 0.1 6 4.3 f 0.9 167f4 504 +_ 12 1.28 & 0.03 5.1 f0.2 47k 1 0.14 f 0.01 46f 1 11.5 f0.3 - - 7722 0.23 f 0.07 1.57f0.12 0.032 f 0.003 0.060 f 0.002 15.5-1 1.3 0.37 k 0.07 0.58 f 0.10 0.33 f 0.22 0.045 k 0.002 0.063 f 0.0 16 0.028 f 0.004 0.096 & 0.004 0.039 f 0.010 - __JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 1073 Results The investigated vegetables contain similar main components but different trace elements.There was no problem in wet ashing by pressure digestion using nitric acid. The composition of the sludge sample is completly different and additional hydrofluoric acid was necessary for its digestion. To check the accuracy of the applied techniques the National Institute of Standards and Technology (NIST) Standard Reference Material SRM 1570 Spinach was analysed. The values determined by the three methods agree with the elemen- tal concentrations recommended by NIST6 within the analyt- ical errors (see Table 2). The analytical errors given include the counting uncertainties and the standard deviations of independent analyses of six separate sample aliquots. In Table 3 the results of an inter-comparison analysis of a sample of spinach are summarized.The determined concen- trations range from 30mgg-' (K) to 20 ngg-' (Ag). In general the agreement between the different analytical tech- niques is excellent. Some experimental limitations are evident with TXRF there are no results for elements lighter than P with NAA it is difficult to detect Pb and often the y-ray lines of Ti and Cu are hidden when a high background of matrix elements occurs. In this analysis the results for K and Ca by ICP-MS are satisfactory. The concentration of Fe is not greater than 0.27 gkg-l and that of Cd is surprisingly high at 3mgkg-'. The analytical errors range from 3 to 30%. The results of TXRF for Ba differ if L or K X-ray lines are used and therefore this uncertainty is increased. B is detected only by ICP-MS P and S only by TXRF and C1 Sc Co and Br only by NAA.C1 and Br are partially lost during digestion. Co was not determined because it was added as an internal standard to the sample. In total 25 elements were determined Table 3 Elemental concentrations in spinach (mg kg-I). Results are means _+ standard deviations (n = 6) by ICP-MS 20 by TXRF and 28 by NAA. It is difficult to obtain good results using ICP-MS for P and S and also for Sc and Se if their concentrations are low. The determined value of Cr by ICP-MS is enhanced by a contribution of Ar and C. On the other hand the detection of trace concentrations below 1 pg g-' is difficult with TXRF as for Ag La Ce and Th without any previous chemical separation. For some trace elements such as Se Pb and U an additional reversed-phase technique has been applied before analysis by TXRF.This involves using dithiocarbamate complexation adsorption of the metal complexes on a Chromosorb column and subsequent elution of the complexes with methanol-chloroform.7 The comparison of the analyses of a sample of cabbage (Table 4) again shows good agreement between the different methods. The range of concentrations is similar but the concentrations of some elements such as Na Mg V Mn Fe Rb and Cd are lower than in spinach and the concentrations of Mo and Pb are higher. Again we obtained no data for the lighter elements using TXRF or for Ti Cu and Pb using NAA. Clearly there are similar limitations to the detection of B P S C1 Sc Co Se Br La Ce and Th using the three different techniques.The number of elements detected and the uncer- tainties are similar to those in the investigation of spinach. A different type of material a domestic sludge sample was also investigated. It contained high concentrations of Ca and Fe (38 and 27 mg g-' respectively). Compared with the aver- age amounts of the elements in the earth's crust high concen- trations of Cu Zn Sc Ag Pb Bi U Cd and Hg are found. The rare earth elements are present at concentrations of 2-70 pg 8-l and U and Th at 39 and 6 pg g - l respectively. In Table 5 all the results for this sample are summarized. More than 40 elements were detected. Most of the results for the three methods agree well with each other including those for Sc and Hg. Some values determined by ICP-MS are lower than those given by the other methods e.g. K Ca Cr As Se Rb and Zr.The probable reason is that the sludge sample had Element B Na Mg A1 P S c1 K Ca sc Ti v Cr Mn Fe c o Ni c u Zn Se Br Rb Sr Mo Ag Cd c s Ba La Ce Sm Hf Hg Pb Th U ICP-MS 41+1 17200 _+ 600 8400 f 200 250 f 50 - - - 32300 f 1600 13300 f 500 22.0 f 2.5 0.65 & 0.02 1.89 f 0.20 7 6 f 2 272 f 8 2.3 f 0.2 11.6 f 0.5 80$3 - - - - 12.1 f0.9 56f2 0.37 f0.04 0.020 f 0.003 2.7 f 0.2 6.7 f 0.4 0.12 f 0.05 0.24 f 0.05 - - - - 0.16 f0.04 0.043 f 0.003 0.14 + 0.04 TXRF - - - - 5000 f 260 4600 k 240 29400 f 1400 15200 + 750 20.9 f 3.6 0.65 k0.15 1.67 f0.51 76+2 281 f 8 2.5 f 0.3 13.1 k0.5 86+3 0.12 + 0.03 13.8 + 0.6 58f2 0.42 f 0.04 3.2 f 0.2 6.2f 1.7 - - - - - - - - - - - 0.19 & 0.05 0.16 f 0.05 - NAA - 17200&800 8600 f 400 310f 15 - - 6760 f 350 29200f1500 12900 f 500 0.052 f 0.002 0.61 f0.14 1.68f0.12 76k4 250f 12 0.35 f0.03 1.9 f0.3 76+3 0.11 fO.O1 33.0+ 1.5 12.2f0.5 4 8 f 3 0.28 f 0.09 0.022 f 0.003 3.0f0.1 0.02 1 f 0.002 6.6 f 0.7 0.19 k 0.07 0.32 f 0.10 0.017 f 0.004 0.028 f 0.003 0.025 f 0.007 0.045 f 0.003 0.14f0.02 - - - Table 4 means f standard deviations (n = 6) Elemental concentrations in cabbage (mg kg-I).Results are Element B Na Mg A1 P S c1 K Ca sc Ti v Cr Mn Fe c o Ni c u Zn As Se Br Rb Sr Mo Cd cs Ba La Ce Pb Th ICP-MS 30.1 f 0.8 570 f 25 2030 f 500 155 f 24 - - - 25500 f 1700 14300 zfr 500 12.5 f 1.7 0.35 $0.05 1.34f0.15 30.5 f 0.9 165 f 13 1.2f0.2 5.5 f 0.2 36.3 f 1.0 - - __ - - 6.0 f 0.2 47.8 f 1.3 1.50 f 0.09 0.14f0.01 10.9 f 0.6 0.10 f 0.02 0.18f0.03 0.85 f 0.05 - - TXRF - - - - 5150 f 580 15600 f 800 32900 f 1500 18500f900 11.2f1.3 0.29 k 0.09 0.92 0.30 29.9 f 1.2 147f5 1.2f0.3 5.6 k 0.2 40.3 f 1.3 0.1 1 f 0.02 6.5k0.3 47.2k 1.5 1.66 f 0.1 8 0.14 f 0.04 9.5 f 1.5 - - - - - - - - 0.94 f 0.07 - NAA - 543 f 25 2030 f 80 165 + 8 - - 8600 k 350 32000 f 1300 15800 f 500 0.024 f 0.001 0.25 + 0.07 1.20 It 0.09 30.7 f 1.2 135f 13 0.10 f 0.01 0.9 f 0.2 35.4f 1.4 0.087 & 0.012 0.11 kO.01 5.8 f 0.2 5.7 f 0.2 41.2 f 2.0 1.29 f 0.07 0.13 f 0.02 0.026 f 0.002 9.2 f 0.9 0.19+0.11 0.27 f 0.15 0.024 + 0.001 - - -1074 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 Table 5 are means f standard deviations ( n = 6) Elemental concentrations in domestic sludge (pg g-'). Results Element Na A1 P S c1 K Ca sc Ti V Cr Mn Fe c o Ni c u Zn Ga As Se Br Rb Sr Y Zr Nb Mo Ag Cd In Sn Sb I Ba La Ce Nd Sm Yb Lu Hf Ta W Pb Bi Th U Mg DY Hg ICP-MS 2070 f 70 5370 + 200 - - - - 41OO& 110 32200 f 2100 72.8 f 1.5 3320 & 70 98f2 166.t6 844f18 25600 f 1000 6.3 + 0.4 74+2 600 f 20 1300f26 6.5 f 0.4 6.0 f 0.3 11.5 f0.6 14.6 & 0.4 220f 13 22f2 240 f 30 50+ 1 4 4 f 1 99k5 12.2 f 0.3 0.3f0.1 70+2 7.0 k 0.4 616+ 16 1 8 f 2 67k2 13.5 k0.9 2.9 f 0.2 3.5 f 0.3 2.4 f 0.2 5.0 k 0.9 3.7k0.1 9.9 + 0.4 4.0 k 0.6 193 f 5 30k 1 5.6 f0.6 38.7 & 1.8 - - - TXRF - - - 18300f 1100 13900 f 700 4870 k 200 38900 & 1100 3080 f 200 105 f 6 200f6 844f 17 27300 f 600 7.3 k 0.7 85+4 646 f 20 1300+26 6.5 f 0.4 8.4 0.7 15.5k0.5 19.6 f 0.9 240+ 12 30f2 337 f 17 56k2 46f 1 105 f 3 13.2 + 0.5 <2 72k2 8.6 k 1.5 613 + 22 21+3 70+8 14k2 - - - __ - - - __ 4.7 + 1.2 5+5 11+3 3.5 f0.6 205 2 5 3 3 f l 6.3 f 0.7 39.1 f 1.0 NAA 2250 f 30 6200 f 190 16800 f 250 - - 3080 f 40 5130f 130 38600fll00 72.5 f 1.2 3230 + 110 97f2 198+4 853 f 17 28400 It 600 6.4f0.1 72f6 636 f 26 1240425 6.9 f 0.8 8.0 f 0.3 15.5 + 0.3 15.1 k0.2 19.0 f 0.7 260k 18 330f 15 3 6 f 3 9 2 f 2 11.7 + 0.9 0.32 f 0.03 8.6 f 0.2 13.9k2.1 661 + 20 22.0 + 0.4 73 f 2 15.3 f 1.5 4.9 f 0.2 4.5 f 0.4 3.2f0.1 1.1 f O .l 7.7 f 0.2 3.4f0.1 11.2f 0.7 3.8 f 0.2 - - - - - 5.8 f O . l 37.8 kO.8 not been completely dissolved. The solution was rather a suspension. This is not a problem for analyses carried out using TXRF but it produces greater variations in ICP-MS and lower mean values. For some elements there is still a demand for an additional reference material.For instance the Hf concentration determined by NAA does not agree with those given by the other methods and needs further investi- gation. Almost all rare earth elements are detectable by ICP-MS. However a comprehensive procedure is required to correct for all the spectral interferences caused by oxides of lighter elements. The lanthanides form particularly strong oxides.8 Conclusions A comparison of the results shows that the accuracy of all three methods is high. The precision depends on the technique and on the concentration. The agreement between the results for the three different methods is good generally better than 10%. The deficiency of light element determinations using TXRF may be overcome by using ICP-MS. NAA is not useful for a large number of samples but it is very important as a reference method to check the quality of the sample preparation. It is very advantageous to apply different techniques to avoid any interference or systematic errors especially for problems that may arise during digestion such as the loss of volatile elements or portions of the sample not being dissolved. References Prange A. and Schwenke H. Adv. X-Ray Anal. 1992,35 899. Klockenkamper R. Knoth J. Prange A. and Schwenke H. Anal. Chem. 1992,64 1115A. NiedergesaB R. Schnier C. and Pepelnik R. J. Radioanal. Nucl. Chem. 1993 168 317. Greim L. Motamedi K. and NiedergesaD R. AKAN-ein Rechenprogramm zur Auswertung bei der Instrumentellen Multielement-Neutronenaktivierungsanalyse GKSS 76JE49 GKSS Research Centre Geesthacht 1976. Sansoni B. in Instrumentelle Multielementanalyse ed. Sansoni B. VCH Weinheim 1985 pp. 3-56. Gladney E. S. O'Malley B. T. Roelandts I. and Gills T. E. NBS Spec. Publ. 260-11 1987. Prange A. Knochel A. and Michaelis W. Anal. Chim. Acta 1985 172 79. Vaughan M. A. and Horlick G. Appl. Spectrosc. 1986 40 434. Paper 4 fO1304B Received March 4 1994 Accepted May 16 1994

ISSN:0267-9477    

DOI:10.1039/JA9940901071

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

JOURNAL O F ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 1075 New High-resolution Inductively Coupled Plasma Mass Spectrometry Technology Applied for the Determination of V Fe Cu Zn and Ag in Human Serum* Luc Moens Peter Verrept and Richard Dams Ghent University Laboratory of Analytical Chemistry Proeftuinstraat 86 8-9000 Ghent Belgium Ulrich Greb Gerhard Jung and Bernd Laser Finnigan MAT Bremen Germany Spectral interferences are a limiting factor in quadrupole inductively coupled plasma mass spectrometry (quadropole ICP-MS). Most of these interferences disappear when a high-resolution magnetic sector mass spectrometer is coupled to the ICP ion source. In this paper results of the first analyses with a new type of a high resolution ICP-MS instrument are shown. The instrument is a commercially available machine (Finnigan MAT Bremen Germany) offering standard resolution settings of 300 3000 and 7000 (MIAM 10% valley definition).With a resolution setting of 3000 V Fe Cu and Zn were determined in a second generation human serum reference material. Human serum diluted 4- to 8-fold was measured. The results expressed as concentrations (pg g-’) [standard deviation (SD) in parentheses] in the freeze dried material for Fe 23.6 (0.8); Cu 10.7 (0.2); and Zn 8.2 (0.8) are in good agreement with the certified values. The very low V content is not certified. In the high-resolution spectrum the V peak was measured next to an approximately 1000 times higher 35Cl’60 peak and a concentration of 0.83 ng g-’ SD 0.09 ng g-’ was found which confirms an earlier radiochemical neutron activation value of 0.67 ng g-’ SD 0.05 ng g-’.For the determination of Ag using the low resolution (300) setting a limit of detection (LOD) of 4.3 pg ml-’ in the solution was found. The instrumental LOD is 10-100 times lower and the experiments show that an investigation of blanks and methods of dealing with memory effects will be necessary before full use can be made of the sensitivity of high resolution ICP-MS. Keywords Inductively coupled plasma mass spectrometry; high resolution; human serum Most of the present instruments for inductively coupled plasma mass spectrometry (ICP-MS) are equipped with a quadrupole. This m/z filter allows positive ions to be separated when their masses differ by one or even half an atomic mass unit. This is usually insufficient to distinguish the atomic ions of the analyte elements from interfering polyatomic ions.The latter consist of atoms originating from the plasma gas (Ar) major constitu- ents of the sample and the solvent and the reagents used to prepare the sample solution. The resulting spectral interferences are an important limiting factor in ICP-MS and their elimin- ation or reduction is a major research topic. Spectral interferences can be diminished by reducing the amount of solvent reaching the plasma,’ by applying electro- thermal vaporization ( ETV),2-6 by chemical elimination of interference generating constituents prior to the measure- ment,7-’1 by selecting appropriate operational12 parameters or by adding substances that change the conditions in the plasma.13-17 Most of the spectral interferences occurring with quadrupole machines disappear with the use of high resolution (HR)- ICP-MS machines which offer a resolution up to M/AM= 10000 (where AM is the mass difference between two singly charged ions of average mass M that produce peaks in the mass spectrum which overlap at 10% of the maximum peak height). The HR-ICP-MS equipment has been available for several years but at a considerably higher price than quadru- pole instruments. Therefore the method has not been widely used for research and few papers have appeared in the l i t e r a t ~ r e . ’ ~ ~ ~ ~ The new instrument called ‘Element’ manufactured by Finnigan MAT (Bremen Germany) was designed from scratch for HR-ICP-MS and will be available for a price that is comparable to the price of the early quadrupole machines.In this work the machine was used for the first time to determine * Presented at the 1994 Winter Conference on Plasma Spectro- chemistry San Diego CA USA January 10-15 1994. trace and ultra-trace elements in a human serum reference material. Measurements were performed on human serum diluted by a factor of 4 or 8 to reduce non-spectral matrix effects. Vanadium Fe Cu and Zn which are hard to determine in serum with a quadrupole instrument due to the presence of severe spectral interferences are measured. When used in the low resolution mode (MIAM = 300) the HR-ICP-MS instru- ment promises limits of detection (LODs) that are up to a factor of 100 lower than those of quadrupole machines.This opens new horizons but at the same time creates the need for a further reduction of the blank levels which for most elements keep the LODs from reaching the level determined by the instrumental noise. This is illustrated for the determination of Ag in the serum reference material. Experiment a1 Instrumentation Measurements were performed with the first Element instru- ment produced by Finnigan MAT. The Element is equipped with a high-resolution magnetic sector mass analyser of reversed Nier-Johnson geometry. It comes with standard resolution settings of 300 3000 and 7000. Since most spectral interferences caused by polyatomic species require a resolution of not more than 2700 the resolving power of the instrument is adequate to solve most interference problems occurring in quadrupole ICP-MS.The ion transmission of the instrument at low resolution (M/AM=300) is high and 1 ppm of In produces a count rate of at least 20 x lo6 counts s-’. With a typical detector background of 0.2 counts s- ’ LODs of the order of 10 pg ml-’ at low resolution are feasible. Several scan modes (magnetic high voltage) are possible one of which is the patented synchro-scan mode in which counteracting and simultaneous electric and magnetic scans make it possible to spend most of the analysis time collecting peak data rather than the background in between peaks.1076 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 The results described in this work were obtained at a time when all the planned features of the instrument had not yet been fully developed on the particular machine used in this study.Some acquisition modes were not yet available. The synchro-scan mode was only available at the low resolution setting (MIAM= 300). When using a resolution of 3000 it was only possible to use the magnet for scanning. It was not possible to peak jump at either the low or the high-resolution settings and only the scanning mode was available. Also some of the software for data acquisition and data reduction was not present on the instrument. This explains why for several experiments non-ideal instrument settings were used in this work. Nevertheless results can be considered typical for the Element instrument. The software required to allow the miss- ing operation modes to be used has been implemented by now and is available on the forthcoming machines.Instrumental conditions are summarized in Table 1. Because the Element HR-ICP-MS machine was only available for a short period of time there was no time for specific optimization of gas flow rates sample uptake rate etc. Therefore compro- mise experimental conditions were chosen that had allowed sensitive accurate and reproducible determinations of 14 trace and ultra-trace elements in human serum with both a Fisons PQ 124 and a Perkin-Elmer Elan 5000 quadrupole machine. Samples and Standards In this work the Second Generation Human Serum Reference Material (RM) was analysed. This material is freeze-dried human serum prepared by Versieck et al. in which trace element concentrations closely approximate those in normal human serum.25 The RM was prepared under rigorously controlled conditions in order to avoid extraneous additions. For several trace elements the concentrations are therefore much lower sometimes by orders of magnitude than in most similar RMs.The material has been analysed by many investi- gators solicited for their established expertise and the concen- trations of 14 elements have been certified.25 The certified concentrations shown in Table 2 are in pg 8-l or ng g-' of the dry material. The corresponding concentrations in pg ml-' or ng ml-I of liquid serum respectively are lower by a factor of 11 than the results shown here. In this work concentrations Table 1 Instrumental conditions R.f. power/kW Ar gas flow rates11 min-' Nebulizer Intermediate Plasma Sample uptake ratelm1 min-' Torch Nebulizer Spray chamber Cones Sampling Skimmer 1350 0.725 0.7 13.5 0.9 Fassel torch (central channel 1.2 mm i.d.) Meinhardt TR-30-A3 concentric Double-pass Scott Type cooled at 5-6 "C Ni 1.1 mm orifice Ni 0.8 mm orifice were determined for Fe Cu and Zn and for Ag and V.The Concentrations of the last two elements are below 1 ng 8-l of dry material and were not certified. Human serum is a rather difficult matrix for ICP-MS. It contains 6-8% of proteins and about 1 % of inorganic compo- nents causing important spectral interferences up to m/z 97. Nevertheless quadrupole ICP-MS has been successfully applied to determine concentrations of 20 trace and ultra-trace elements in human serum without any sample preparation except for a 5- to 10-fold dilution with 0.14 mol I-' HN03.24 Interference factors were determined experimentally and were defined as the apparent concentration in pg I-' of the analyte element caused by a 1 g 1-' concentration of the interfering element.26 These interference factors were used in this work to estimate the magnitude of spectral interferences.For determinations with HR-ICP-MS sample preparation was limited to reconstitution of the serum and subsequent dilution with 0.14 mol 1-1 HNO,. Samples were prepared under clean room conditions. For the determinationss of Ag and V the serum was diluted by a factor of approximately 4 for the other elements by a factor of about 8. Three independent sample preparations were performed and the solutions were measured immediately after the preparation of the samples.Commercial standard solutions (Atomic Absorption Standard Solutions Janssen Chimica B-2440 Geel Belgium) at a concentration of 1 g 1- were used to prepare appropriate standards. Internal standard Co for the determination of Fe Cu Zn V or In for the determination of Ag was added at the 10 ng ml-' level to the samples and at a level of 1 ng ml-' to the standards. Standards were diluted with 0.14 mol 1-1 HN03. For all preparations 18 Mi2 Milli-Q water (Millipore Bedford MA USA) and HNO purified by sub-boiling in quartz equipment was used. Acquisition For the determination of Ag the low resolution setting (MIAM = 300) and the synchro-scan acquisition mode were used for standards blanks and samples. The 105-125 m/z range was scanned and data were acquired at each integer mass.For the three independently prepared sample solutions and for all standards five scans were performed taking 30s each. The blank level was determined for 15 consecutive scans. The high resolution mode (M/AM=3000) was used for the measurements of Fe Cu Zn and V. Iron Cu and Zn were determined in the serum solution diluted 8-fold. Three consecu- tive scans were performed for each of the mass ranges of interest. Thus first three scans were performed over the Fe mass range (55.90-55.99 u) then three scans over the Cu mass range (62.91-62.99 u) the Zn mass range (65.90-65.97 u) and finally over the Co mass range (58.94-59.00 u). A similar procedure was applied for the determination of V in the serum diluted 4-fold scanning the magnet over the V (50.94-51.00 u) and Co m/z ranges.Each scan took 15.8 s and allowed data to be acquired at 40 different masses. Blanks standards and samples were measured in that order. Obviously the data acquisition procedure can be improved. The software that is available on the instrument can now select Table 2 Certified concentrations and 95% confidence intervals in the Second Generation Freeze-Dried Human Serum RMZ5 Content 20.2 0.76 7.7 25.9 3.6 11.1 9.6 95 Yo confidence interval 17.5-23.3 0.67-0.87 7.4-8.0 24.4-27.4 3.0-4.2 10.7-11.5 9.4-9.8 Content 19.6 1.05 48.8 1.85 7.5 2.0 10.0 95 % confidence interval 15.6-23.6 1 .00- 1.10 45.0-52.6 1.52-2.1 8 6.7-8.3 1.7-2.5 7.7- 12.3JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 1077 36 an operation mode that allows fast electric scanning making the use of an internal standard more effective. - - Results and Discussion Silver was determined in the human serum RM in order to test the possibility of determining extremely low concentrations of elements in the low resolution mode. The concentration of Ag was expected to be below 1 ng g-' in the dry material a value determined with radiochemical neutron activation analy- sis (RNAA)27. This corresponds to about 80pgml-' in the re-constituted serum and 20pgml-' in the serum diluted 4-fold. This is below the practical LODs that are attainable with quadrupole ICP-MS as was experienced in previous Fig. 1 shows the results of 15 consecutive blank measure- ments. The blank turned out to be relatively high namely 38.5 pg ml-' with an SD of 1.4 pg ml-'.Using the 3s criterion an LOD of 4.3 pg ml-' could be obtained. This is much higher than what is theoretically possible for HR-ICP-MS but can easily be explained by the high blank value caused by memory effects due to the use of highly concentrated Ag solutions in testing and optimization experiments performed on that par- ticular instrument. This example clearly shows that the extreme sensitivity at low resolution of magnetic sector instruments may be an incentive for a new round in the battle against blanks. It can indeed be expected that for many elements LODs will be imposed by the blank level and not by the instrumental background. If blanks can be controlled however the sensitivity of the machine can be used in analytical situ- ations presently requiring analyte enrichment or for element sensitive detection after chromatographic separation for trace element speciation.When measuring the Ag peak at m/z 107 a concentration of 0.59 ng g-' for the freeze dried material was found with a standard deviation of 12%. For Ag at m/z 109,0.47 ng g-' was found with an SD of 19%. These values are substantially lower than the only value ever determined for the material (0.93 ng g-l) by RNAA.27 In quadrupole ICP-MS the 56Fe-isotope (91.7% abundance) with a mass of 55.935 u suffers from an interference by ArO (55.957 u). The magnitude of the interference strongly depends on the experimental conditions and when serum diluted %fold is measured it is equivalent to an Fe content of 1.7-3.3 pg 8-l in the dry serum. There is an additional interference by CaO (55.957 u) due to the high Ca content of serum.Using the interference factors determined experimentally mentioned before the apparent Fe concentration due to CaO can be estimated to be about 0.63 pg g-' for dry serum. Compared with the actual Fe content of 25.9 pg g-' these interferences are moderate but it is nevertheless preferable to measure 54Fe 40 r 39 E g 38 I - 23 9 3 7 ' Mean - - 1 Fig. 1 Result of 15 consecutive scans of a blank solution. The mean apparent Ag concentration is 38.5 pg ml-' (SD 1.4 pg m1-I) and the detection limit is 4.3 pg ml-' 1600 I I 1400 7 1200 In In c 5 1000 8 '& 800 '5 600 5 400 200 0 7 \ c > a c 55.89 55.9 55.91 55.92 55.93 55.94 55.95 55.96 55.97 55.98 m/z Fig.2 HR-ICP-MS spectrum at m/z 56 of human serum diluted 8-fold; resolution = 3000 (5.8% abundance) for which interferences caused by ArN and ClOH are less important.The 56Fe peak can be measured free of interference when a resolution of 2500 or more is available. In Fig. 2 part of the HR-ICP-MS spectrum (resolution of 3000) of serum diluted 4-fold is shown. The 56Fe-peak is clearly separated from the peaks for ArO and CaO. The Fe content of 23.6 pg g-I (SD 0.8 pg g-') for the dry material determined with HR-ICP-MS is in acceptable agreement with the certified value of 25.9 1.5 pg 8 - l . When analysing human serum with quadrupole ICP-MS interferences occur for both Cu isotopes as shown in Table 3. Using the experimental interference factor the apparent Cu concentration in the dry serum due to ArNa causing an interference at m/z 63 can be estimated to be about 9.46 pg g-'.Comparison with the actual concentration (11.1 pg g-') shows that this interference is too large to be appropriately corrected for. Determination via m/z 65 is possible since the apparent Cu concentration is only about 7% of the actual content. With a resolution of 3000 both interferences can be separated from the analyte peak. The Cu concentration was determined uia the peak at m/z 63. A value of 10.7 pg 8-l with an SD of 0.2 pg g-' was found for the dry material which is in excellent agreement with the certified value of (11.1 k0.4) pg g-'. The determination of Zn in serum with quadrupole ICP-MS is hampered by interferences caused by S-containing polya- tomic ions for all five isotopes. In practice 66Zn (27.9%) and especially 68Zn (18.8%) can be used if correction procedures e.g.using matrix simulating blanks,29 are applied. For HR-ICP-MS the 66Zn isotope was measured for which quadru- pole ICP-MS shows interferences by 34S1602 33S16021H and 32S'60180 accounting for an apparent Zn content of around 0.6pgg-' in dry serum. All interferences can be separated from the analyte signal when using a resolution of 3000. For the dry material a concentration of 8.2 pg g-' with an SD of 0.8 pg g-' was found which is in reasonable agreement with the certified value of (9.6 50.2) pg g-'. Table 3 Spectral interferences in the determination with quadrupole ICP-MS of the Cu content in the Second Generation Human Serum RM Apparent Cu concentration in human Species Mass M/AM serum/pg g-' 6 3 C ~ (69.2%) 62.93 40Ar23Na 62.95 2790 9.46 65Cu (30.8%) 64.93 2 64.97 1548 0.80 32s160 1H1078 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 300 1 250 r 'v) 200 v) 2 c 8 150 \ c z v) .- g 100 4- - 50 n i 50.918 50.928 50.938 50.948 50.958 50.968 50.978 m/z Fig.3 HR-ICP-MS spectrum at m/z 51 of human serum dilated 4-fold resolution = 3000 The determination of the V concentration in human serum is a difficult problem. The concentration of V is low V is almost mono-isotopic ("V 99.7% 50.944 u) and there is a strong interference from 35Cl'60 (50.964 u) and 37C114N (50.969 u). For the human serum RM used in this work only one concentration value has been reported Byrne and Versieck3' used RNAA and found a value of 0.67 ng g-' SD 0.05 ng g-l) in the freeze dried material.Using the experimen- tally determined interference factor the apparent concentration of V due to the interference by 35C1160 can be calculated to be 1300ngg-' which is more than 3 orders of magnitude higher than the actual content. Evidently quadrupole ICP-MS does not allow the V content to be measured in this material. To obtain a calibration curve for V standards with a content of 10 pg ml-' 100 pg ml-' 1 ng ml-' and 10 ng ml-' were measured. The V content of the blank was 2.5 pgml-'. The calibration curve is linear with a correlation coefficient of nearly 1 a slope of 0.08701 mlng-1 and an intercept of O.ooOo28. In the high-resolution (3000) spectrum shown in Fig.3 the V peak is very small compared with the C10 peak the top of which is at about 72000 counts s-'. Two indepen- dent determinations were performed. The average concen- tration found was 0.83 ng g-' for the dry material with an SD of 0.09 ng g-'. The agreement with the result using RNAA is satisfactory. Conclusions The new commercial HR-ICP-MS instrument by Finnigan MAT (Bremen Germany) allows the concentrations of Fe (uia 56Fe) Cu (via 63Cu) and Zn (uia 66Zn) to be accurately and precisely determined in human serum. Sample preparation can be limited to 8-fold dilution and spectral interferences experi- enced with quadrupole ICP-MS are eliminated by using a resolution setting of 3000. In serum diluted 4-fold the 51V peak was measured next to a 1000 times higher peak of 35C1160.The content of V found of 0.83 ng g-' (SD 0.09 ng g-') in the solid material is confirmed in the literature. When used in low resolution mode (300) the instrument is characterized by LODs of the order of 10pgml-'. Limits of detection for Ag turned out to be substantially higher due to memory effects. A wider use of HR-ICP-MS can be expected to start new research to further reduce blank and memory effects. This will be necessary to allow the extremely low instrumental detection limits to become useful. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 References Alves L. C. Wiederin D. R. and Houk R. S. Anal. Chem. 1992 64 1164. Gregoire D. C. J. Anal. At. Spectrom. 1988 3 309. Tsukahara R. and Kubota M.Spectrochim Acta Part B 1990 Hulmston P. and Hutton R. C. Spectrosc. Int. 1991 3 35. Carey J. M. Evans E. H. Caruso J. A. and Shen W.4 Spectrochim. Acta Part B 1991 46 1711. Carey J. M. and Caruso J. A. Crit. Rev. Anal. Chem. 1992 23 397. McLaren J. W. Mykytiuk A. P. Willie S. N. and Berman S . S. Anal. Chem. 1985 57 2907. Plantz M. R. Fritz J. S. Smith F. G. and Houk R. S. Anal. Chem. 1989,61 149. Sheppard B. S. Shen W.-I. Caruso J. A. Heitkernper D. T. and Fricke F . L. J. Anal. At. Spectrom. 1990 5 431. Goossens J. and Dams R. J. Anal. At. Spectrom. 1992 7 11 67. Goossens J. Moens L. and Dams R. J. Anal. At. Spectrom. 1993 8 921. Goossens J. Vanhaecke F Moens L. and Dams R. Anal. Chim. Acta 1993 280 137. Evans E. H. and Ebdon L. J. Anal. At. Spectrom.1989 4 299. Evans E. H. and Ebdon L. J . Anal. At. Spectrom. 1990 5 425. Lam. J. W. and Horlick G. Spectrochim. Acta Part B 1990 45 1313. Branch S. Ebdon L. Ford M. Foulkes M. and ONiell P. J. Anal. At. Spectrom. 1991 6 151. Beauchemin D. and Craig J . M. Spectrochim. Acta. Part B 1990 46 603. Bradshaw N. Hall E. F. H. and Sanderson N. E. J. Anal. At. Spectrom. 1989 4 801. Morita M. Ito H. Uehiro T. and Otsuka K. Anal. Sci. 1989 5 609. Kim C.-K. Seki R. Morita S. Yamasaki S.-I. Tsumura A. Takaku Y. Igarashi Y. Yamamoto M. J. Anal. At. Spectrom. 1991 6 205. Walder A. J. and Freedman P. A. J . Anal. At. Spectrom. 1992 7 571. Walder A. J. Platzer I. and Freedman P. A. J. Anal. At. Spectrom. 1993 8 19. Walder A. J. Koller D. Reed N. M. Hutton R. C. and Freedman P. A. J. Anal. At. Spectrom. 1993 8 1037. Vanhoe H. J. Trace. Electrolytes Health Dis. 1993 7 131. Versieck J. Vanballenberghe L. De Kesel A. Hoste J. Wallaeys B. Vandenhaute J. Baeck N. Steyaert H. Byrne A. R. and Sunderman F. W. Jr. Anal. Chim. Acta 1988 204 63. Vanhoe H. Goossens J. Moens L. and Dams R. J. Anal. At. Spectrom. 1994 9 177. Lin X Van Renterghem D. Cornelis R. and Mees L. Anal. Chim. Acta 1988 211 231. Vanhoe H. Ph.D. Thesis 1992 Gent. Vanhoe H. Vandecasteele C. Versieck J. and Dams R. Anal. Chem. 1989 61 1851. Byrne A. R. and Versieck J. Biolog. Trace Elem. Research. 1990 27 529. 45 779. Paper 4/01 178C Received February 28 1994 Accepted May 23 1994

ISSN:0267-9477    

DOI:10.1039/JA9940901075

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 1079 Analysis of Liquid-phase Tungsten Hexafluoride Residue by Inductively Coupled Plasma Mass Spectrometry With Ultrasonic Nebulization* Robert L. Sutton Airco Electronic Gases P.O. Box 72338 Research Triangle Park NC 27709 USA A method was refined for the determination of trace elemental contaminants in high-purity tungsten hexafluor- ide. The procedure is based on the collection of a liquid-phase sample of WF6 generating a residue by evaporation of the sample and subsequent digestion in concentrated ammonia solution. After sample digestion and workup analysis was performed by inductively coupled plasma mass spectrometry (ICP-MS) with ultrasonic nebulization. The reliability of the analysis method and instrument limits of detection were found to be superior to ICP-MS with a pneumatic nebulizer and water-cooled spray chamber.The following limits of detection (in ng C M - ~ ) were observed Na 0.2; Mg 0.01; P 3.2; K 1.5; Cr 0.02; Fe 0.5; Ni 0.03; Cu 0.08; As 0.07; Pb 0.01; Th 0.02; and U 0.03. Keywords Inductively coupled plasma mass spectrometry; ultrasonic nebulization; tungsten hexafluoride; liquid-phase residue analysis; cylinder liquid-phase sampling The drive by semiconductor manufacturers towards smaller device geometries in the production of integrated circuits has fuelled the desire to obtain process chemicals with minimal amounts of impurities. Chemical and speciality gas suppliers face an ever increasing demand by the semiconductor industry for decreased amounts of impurities in their products as seen in customer specifications.Tungsten hexafluoride (WF,) is primarily used by the semi- conductor industry as a refractory interconnecting material between areas on an integrated circuit. The use of WF has a number of advantages among which are (i) low resistances of tungsten and its associated silicide (WSi,); (ii) the thermal expansion coefficients for W and WSi closely match the value for silicon; (iii) the migration of tungsten on the chip surface is relatively low; and (iv) the ability to be selectively deposited onto semiconductor material1 The WF is deposited onto the surface of a wafer by means of chemical vapour deposition and can be reduced by either hydrogen or silicon in the following manner 2WF,(g)+ 3Si(g)-.2W(S)+3SiF,(g) ( 1 ) WF6(g)+ 3H2(g)+W(s)+6HF(g) (2) or The analysis of WF for trace elemental contaminants by inductively coupled plasma mass spectrometry (ICP-MS) has traditionally been very difficult.Owing to the large amount of W present signal suppression is a primary concern. Streusand et aL3 reported the inability to analyse WF gas directly by injecting it into the plasma and by direct analysis of the gas bubbled into water. Although analysis of WF was accomplished by collecting a liquid-phase sample of WF evaporating the liquid using a nitrogen flow and working up the residue in either water or hydrofluoric acid concern was expressed that volatile metallic impurities would be lost when the sample was evaporated. To that end a separation procedure involving extraction of an aqueous sample of WF with CI- benzoin oxime to remove the tungsten selectively was detailed.A similar method for extracting tungsten from an aqueous sample of WF using a lipophilic amine (trioctylamine) or a lipophilic quaternary ammonium salt (Adogen 464) was described by Saab et aL4 There is a continuing argument within the semiconductor industry about the best method for analysing gas products * Presented at the 1994 Winter Conference on Plasma Spectro- chemistry San Diego CA USA January 10-15 1994. such as WF for trace elemental contaminants. Although semiconductor manufacturers use WF in the gas phase utiliz- ing extraction techniques for the analysis of a gas-phase hydrolysis sample can lead to contamination concerns even in a clean room setting.In addition the time required to perform multiple extractions can be extensive. Sampling WF in the liquid phase and analysis of the resultant residue offers the supplier and customer a 'worst-case scenario' for the extent of trace elemental contaminants within the cylinder. Recent experiments performed in our laboratory investigating gas- phase hydrolysis and liquid-phase residue chlorine analysis by ICP-MS showed that the level of trace elemental contaminants remains relatively constant in both phases to 90% depletion of product within the ~ylinder.~ For laboratories that routinely analyse WF6 liquid-phase residue samples by ICP-MS signal suppression due to the tungsten substrate is still a major concern if using a pneumatic nebulizer and conventional water-cooled spray chamber.To illustrate this problem a graph showing the general decline in the average counts per ppb (25 ppb of In as the internal standard) for the semiquantitative analysis of nine WF6 liquid- phase residue samples is given in Fig. 1. For the last five samples analysed over 97% of the analyte signal is depressed compared with the solvent blank analysed at the beginning of this set of samples. Preconcentration of the residue sample prior to analysis or sample dilution prior to analysis to reduce the amount of tungsten present were not considered as viable options because a number of trace elemental contaminants are 4000 I 1 3500 &q R I I ~ 3000 2 2500 L v) 4- 5 2000 & 1000 0 3 1500 0 .- v) 500 n " Blank 100ppb ' Y check Liquid-phase residue samples standard Fig.1 titative analysis of WF6 using a pneumatic nebulizer Internal standard counts per ppb; comparison for a semiquan-1080 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 - required to be in the sub-ppb range. To this end analyses of WF liquid-phase residue samples were attempted using ultra- sonic nebulization followed by sequential heating and cooling of the generated aerosol in order to take advantage of the increased sample intake into the plasma. Although signal depression due to the tungsten substrate would not be elimin- ated it was hoped that the increase in signal intensity for analytes of interest would be sufficient to allow for reliable quantitative analysis of WF liquid-phase residues. h l.4 v4 Experimental" Sampling All liquid-phase sampling of WF was performed at the Airco Electronic Gases facility in San Marcos CA USA after the product had been filled into specially designed cylinders out- fitted with appropriate cylinder valves dependent on customer specifications.The WF sampling apparatus is shown in Fig. 2. The WF6 cylinder was secured into a cylinder inverter so that a liquid- phase sample could be collected. Except for the metal CGA connector to the cylinder all parts of the sampling system were made from Teflon or other inert plastic material. Grade 6 (or 99.9999%) N2 was used as the system purging gas. V1 V2 and V4 are two-way stopcocks with connectors for 1/4in Teflon tubing while V3 accomodates 1/8 in tubing (Galtek Chaska MN USA). Between V2 and V3 is a 1/4-1/8in reducing union (Fluoroware Chaska MN USA) to allow the sample to enter a pre-leached 100 ml polypropylene calibrated flask (Nalgene Rochester NY USA).The connection of the sample line to the calibrated flask is illustrated in Fig. 3. A hole was tapped into the cap of a flask in order to accomodate a 1/8-1/4 in reducing union and a 1/4 in T joint (Plasmatech El Monte CA USA). A 1/8 in Teflon line was threaded through the connector system to allow the sample to enter the flask and to allow for the residual gases to leave the flask to the disposal system. The following are details of the method for collecting a sample from a given WF cylinder (i) attach the inverted WF6 cylinder to the sampling system; (ii) attach a 250ml Nalgene calibrated flask to the sampling system for purging; (iii) open V1 V2 V3 and V4 and purge the,system with Grade 6 N2 at 10 psig; (iu) close V1 and v3; (u) slowly open the WF cylinder valve and open V3 (sample should start flowing into the flask); (ui) at room temperature fill the flask with approximately 200 g (or 58 ml) of WF,; (vii) close the WF cylinder valve and close V2 and V3; (uiii) open V1 V2 and V3 to allow the residual WF6 in the system to flow into the flask and allow * The sampling of any gas under pressure in an approved cylinder always contains an element of risk.In addition the toxic and corrosive nature of WF multiplies the inherent dangers of sampling. It is very strongly recommended that sampling of gas- or liquid-phase WF (or any similar product) be performed only by personnel specially trained in these procedures in a suitable environment with proper safety and disposal equipment.Sample in 1 Plasmatech No. 16F424 screws into flask cap VOL. 9 * To scrubber in Fur& 2-way line Fig. 3 Connection of WF6 sampling line to sampling flask line to purge with N,; (ix) close V2 V3 and V4; (x) remove flask from sampling system and cap immediately; ( x i ) attach a tared 100 ml Nalgene calibrated flask to the sampling system; (xii) repeat steps (iii)-(x) in order to collect a liquid phase sample; and (xiii) weigh the calibrated flask with the sample in order to determine sample mass. The sample evaporation system is shown in Fig.4. A flow of 2-3 psig of Grade 6 N was sent through each flask for approximately 24 h or until only WF residue was present in the sample flask.Additional flasks containing WF may be placed in positions 2 through 5 thus allowing for additional samples to be evaporated simultaneously. The flask (or flasks) was (or were) removed from the evaporation apparatus capped and sent out for analysis. Reagents and Sample Preparation After the liquid-phase residue had been generated the sampling flasks were shipped to the Electronic Development Facility (EDF) laboratory of Airco Electronic Gases in Research 1 2 3 4 5 To scrubber line r- I WF sample T oil vapour trap High purity N Fig. 2 Liquid-phase metals sampling apparatus for WF,; V1-V are two-way stopcocks Fig. 4 WF sample evaporation manifoldJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 1081 Triangle Park NC USA for analysis.The laboratory is part of a Class 10000 clean room facility. Depending upon the needs of the San Marcos facility and the analysis load of the EDF laboratory samples were usually analysed within two days of receipt. Before any analyses were performed two sets of solutions were prepared using single element National Institute of Standards and Technology (NIST) traceable 1000 ppm stock standards (Inorganic Ventures Lakewood NJ USA). One solution contained 10 ppm each of Be In and Bi (used as the internal standards blend). The other solution (used for preparation of calibration standards and hereafter referred to as the intermediate stock standard solution) con- tained the following elements at 10 ppm each B Na Mg Al P K Ca Ti Cr Mn Fe Ni Co Cu Zn Ga As Mo Sn Sb Pb Th and U.Both stock solutions were prepared in a matrix of 5% ultrapure HNO (Seastar Chemicals Sidney BC Canada) and de-ionized H,O and stored in pre-cleaned 125 ml PTFE bottles with a shelf life of three months. For each sampling flask 5 ml of ultrapure ammonia solution (Seastar Chemicals) was added the flask was capped the ammonia solution swirled around the interior surface of the flask and the sample allowed to digest for a minimum of 1 h to release the elemental contaminants from the WF6 on the walls of the flask. After the sample had digested approximately 50ml of ultrapure de-ionized water was added to each flask. A 250pl volume of the internal standards blend (resultant concentration of 25 ppb for each internal standard) was pip- etted into each flask.Each sample flask was filled to the mark with ultrapure de-ionized water capped and mixed several times. Samples can be worked up several hours in advance of analysis but it is recommended that analysis occurs as soon after preparation as possible. When ready for analysis an aliquot of each flask was poured into two separate pre-leached 15 ml cuvettes (Becton Dickinson Lincoln Park NJ USA) Table 1 Instrument conditions and operating procedures ICP torch Argon flow rates/l min-' Outer Auxiliary Aerosol carrier Sample flow rate/ml min-' Forward power/kW Operating pressures/Torr* Interface Intermediate Analy ser Instrument calibration Standard VG torch 13.5-14.5 0.5- 1 .O 0.70-0.80 regulated by mass flow 1.0 typical 1.3 typical controller 1.1-1.4 < 10-4 1.1-1.5 x Mass response and dual detector calibrations performed daily per laboratory SOP before analysis of samples * 1 Torr 133.322 Pa.Table 2 Instrument analysis conditions and placed into an autosampler rack (Gilson Medical Electronics Middleton WI USA). Calibration standards were prepared by filling four labelled 100 ml Class A calibrated flasks approximately one-half full with ultrapure de-ionized water. Ultrapure ammonia solution ( 5 ml) was added to each flask the flasks were capped and the contents agitated. Using the intermediate stock standards solution calibration standards of 25 50 and 100ppb were prepared (one flask was used as the solvent blank). Next 250 p1 of the internal standard blend was pipetted into each of the four flasks (resultant concentration of 25ppb of each internal standard).The flasks were filled to the mark with ultrapure de-ionized water capped and the contents agitated. Finally the contents of each flask were transferred into separ- ate labelled 125 ml high density polyethylene rectangular bottles and placed into a dedicated autosampler rack. As with the residue samples the calibration standards were prepared on the day of analysis. Instrumentation All analyses were performed on a VG Elemental PlasmaQuad Turbo I1 Plus ICP-MS instrument (Cheshire UK) outfitted with a Cetac Technologies U-5OOOAT ultrasonic nebulizer (Omaha NE USA) operating with the heated and chilled regions at 140 and 4"C respectively. Instrumental parameters are presented in Table 1. Data Acquisition The ICP-MS instrument was tuned to the signal of "'In before instrument calibration was attempted to achieve a response of 80 000-120 000 counts on the ratemeter.The instrument can be tuned to a higher response rate but at a cost of increased wear on the electron multiplier in the detection system. After instrument calibration had been performed analysis element procedure and method files were run according to the con- ditions in Table 2. Data were recorded in real time and processed on a Compaq Deskpro 386/20e running the PQVision instrument control software under IBM OS/2 version 2.0. All quantitative analyses were performed with no background (or solvent blank) correc- tion so that a four point calibration curve (0 25 50 and 100 ppb for each isotope of interest) could be generated.All semiquantitative analyses were performed with background correction for the solvent blank. For this work the detection limit is defined as the amount of the element necessary to give a response equal to three times the standard deviation of the mean concentration of the 25 ppb calibration standard (for quantitative analyses) after the four point calibration curve had been generated for each element of interest. Parameter Detector mode Data acquisition mode Number of scans per sample Rinse time between sample+ Sample uptake delay/s Isotopes monitored (m/z) Internal standards (rn/z) Internal standard concentration (ppb each) Dwell time/ps Measurements per mass peak Quantitative analysis Dual (analogue and pulse counting) Peak jumping 5 at 40 s each 90 90 9Be llB 23Na 24Mg 27Al 31P 39K 44Ca 48Ti 52Cr s5Mn 56Fe 58Ni 59Co,65Cu 66Zn 69Ga "AS 9 5 M ~ '"In "OSn lZISb 208Pb 209Bi 232Th and 238U 9Be 'lsIn and '09Bi 25 10240 per isotope 5 Semiquantitative analysis Dual Default continuous scan from m/z 6-245 1 at 60 s 90 90 Those listed for quantitative analysis plus minus the region of m/z 180-190 any others of interest 1151n 25 320 per channel1082 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL.9 Results and Discussion Signal Suppression Streusand et d3 reported in their initial investigations of hydrolysed WF analysis by ICP-MS that the amount of signal suppression precluded analysis for trace constituents although analysis of the residue was possible (a view shared by Saab et d4).Before the USN was acquired for routine analyses all WF residue samples were digested in concentrated NH,OH followed by the addition of HNO to reduce the pH of the sample to approximately 2. This procedure could not be used with the USN because of the likelihood of salts plugging up the injector of the torch so the addition of HN03 to the digested sample was discontinued. However there was concern that some of the W in the residue would enter into solution by complexing with the NH,. A plot of the relative counts of the internal standards used during a typical quantitative analysis run of WF is shown in Fig. 5. The initial mean counts for the solution blank for 'Be "'In and 'O'Bi were approximately 30 000 35 000 and 125 000 respectively. From looking at the graph it is evident that the signal intensities of the internal standards have increased for the calibration standards and the re-analysed 25 ppb cali- bration sample (used for detection limit calculations) compared with the solution blank.The only apgreciable drop occurred in the sample marked R0533 which is an unpurified sample of WF6 (approximately 43% of the solution blank level for 'Be). However the relative count rate for the internal standards increased for the final sample (reanalysis of the 100ppb calibration standard). When using a pneumatic nebulizer a signal drop of 70% for all internal standards was not uncom- mon and would continue with additional samples of WF,. A similar trend was noticed for the semiquantitative analysis of the samples using a pneumatic nebulizer.The increase in the internal standard counting rates for the standards was quite unexpected and not an isolated incident (although other quantitative analyses did not show as pro- nounced an increase in internal standard count rates). The only difference in the calibration standards (aside from the obvious increase in analyte concentration) is that the amount of HNO increases as one progresses from the solvent blank to the 100ppb calibration standard (since the stock solution is in a 5% HNO matrix). An experiment was performed to investigate the effect of adding microlitre amounts of a 5% HNO solution to the solvent blank and lower concentration calibration standards so that the amount of 5% HNO added to each calibration standard (whether from a blank solution or from the stock intermediate calibration solution) were equivalent. Figs.6-8 show the internal standard count ratios for four sets of calibration standards prepared in 5% NH40H without HNO matching 5% NH40H with HNO matching 10% NH,OH without HN0 matching and 10% NH40H with In TI 700 600 1 / I 160 D1 60 20 0 25 50 Standard solutions/ppb 100 Fig. 6 Comparison of relative signal intensities of beryllium for a typical WF quantitative calibration A 5% NH40H and no matrix matching; B 5% NH,OH with matrix matching; C 10% NH,OH and no matrix matching; and D 10% NH40H with matrix matching A 1 300 r A Ir 5 0 1 1 I I I J 0 25 50 100 Standard solutions/ppb Fig. 7 Comparison of relative signal intensities of indium for a typical WF6 quantitative calibration (see legend of Fig.6 for conditions) D 0 25 50 Standard so I u t i o ns/ p p b 100 Fig. 8 Comparison of relative signal intensities of bismuth for a typical WF quantitative calibration (see legend of Fig. 6 for conditions) HNO matching. The stability in internal standard counting rates for Be and In was enhanced by matrix matching with HNO and by increasing the concentration of NH40H to 10%. The stability of Bi was relatively unchanged except for the case of 10% NH40H with matrix matching where a degradation in signal intensity was noted. It is apparent that the dynamics of the plasma interactions using this particular matrix warrant further investigation. A recent paper by Zhu6 detailing a membrane desolvation system to remove volatile solvent components between the nebulizer and the torch injector may lead to further increases in signal stability for this type of analysis.1 I I 1 I I I I Fig. 5 Relative signal intensities of internal standards for a typical quantitative analysis of WF Analysis of Samples and Detection Limits All samples submitted for analysis were routinely run in the quantitative mode first and re-run in the semiquantitative mode as outlined in Table 2. Correction for internal standard drift in the quantitative mode was performed using an interpol- ative procedure as opposed to assigning one internal standardJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 1083 Table 3 Trace elemental contaminants in high purity WF Element Boron Sodium Magnesium Aluminum Phosphorous Potassium Calcium Titanium Chromium Manganese Iron Nickel Mean concentration/ ng cm-3 1.59 6.92 0.55 2.34 3.77 4.07 3.48 0.23 0.80 0.47 2.89 11.4 Mean concentration/ Element ng cmP3 Cobalt Copper Zinc Gallium Arsenic M ol y bdenum Tin Antimony Lead Thorium Uranium 0.38 7.23 1.04 0.06 0.36 0.25 0.11 0.17 0.03 0.03 28.1 Table 4 Representative detection limits for WIF quantitative analysis Detection limit/ Element ng cm-3 Boron Sodium Magnesium Aluminum Phosphorous Potassium Calcium Titanium Chromium Manganese Iron Nickel 0.20 0.22 0.0 1 0.07 3.18 1.50 2.05 0.01 0.02 0.0 1 0.55 0.03 Detection limit/ Element ng cm-3 Cobalt Copper Zinc Gallium Arsenic Molybdenum Tin Antimony Lead Thorium Uranium 0.01 0.08 0.09 0.05 0.07 0.01 0.01 0.01 0.0 1 0.02 0.03 a regular basis.The recent introduction of a commercial membrane desolvation system should further reduce the amount of common interferences listed above.Detection limits were routinely calculated for each quantitat- ive analysis performed as outlined previously and are presented in Table 4. Concern has been raised in the past about volatile species escaping from the sample flask (notably Cr in the form of Cr0,F2).4 Earlier recovery studies conducted using the pneu- matic nebulizer on WF samples spiked prior to sample introduction and residue generation using a multi-element standard containing all of the elements listed in Table 3 ranged from 70% to 133% recovery for a 100 ppb spike (three sets of five spiked samples submitted over a period of three months). Semiquantitative analyses on the residue samples were also performed to attempt to collect data on 42 additional elements.It was not possible to obtain reliable data on 181Ta '''Re and IE7Re because of the interferences from the major mass peaks of W (m/z 182 183 184 and 186). In addition it was found that significant peaks were generated by the formation of WO+ which interfered with the mass peaks for Hg. No additional elements were detected in any of the samples above background levels. The author wishes to thank Patricia Clarke of the San Marcos CA USA facility and her staff for their invaluable assistance in the collection and preparation of the WF samples that were analysed in this paper and Brian Zievis of the RTP NC USA facility for his technical collaboration with the analysis procedures.References for each element (except for 232Th and 238U which were referenced directly to 209Bi). Typical correlation coefficients for each four-point calibration curve were 0.9990 or higher for all elements. A series of nine high purity WF residue samples were submitted over a period of four months in 1993 for analysis by ICP-MS. The mean concentrations for 23 elements deter- mined in the quantitative mode are shown in Table 3. Although the USN increases instrument sensitivity and greatly reduces the number of interferences some still exist but are compen- sated for by matrix matching the calibration standards and samples. The ions NOH+ and 31P ArH' and 39K CO,' and 44Ca ArNH' and 55Mn and ArO' and ',Fe are of concern because of the NH and H,O present in all of the standards and samples. In addition attaining low background levels for Na and B was possible if the glassware was rinsed (in the case of the nebulizer) or acid cleaned (for the torch) on 1 Wolf S. and Tauber R. N. Silicon Processing for the VLSI Era Lattice Press Sunset Beach California 1986 vol. 1 pp. 399-405. 2 Broadbent E. K. and Ramiller C. L. J. Electrochem. SOC. 1984 131 1427. 3 Streusand B. J. Yost B. E. Govorchin S. W. Fry R. C. Padula F. J. and Hughes S . K . Materials Research Society Symposium Proceedings on VLSI/ULSI Applications V Materials Research Society Pittsburgh 1990 pp. 251-5. 4 Saab W. Sarda A. and Cote G. Anal. Chim. Acta 1991 248,235. 5 Sutton R. L. Zievis B. and Conroy M. paper 1098 presented at the 44'h Pittsburgh Conference and Exposition on Analytical Chemistry and Applied Spectroscopy Atlanta GA USA March 6 Zhu J. J. paper TP7 presented at the 1994 Winter Conference on Plasma Spectrochemistry San Diego CA USA Janaury 10-15 1994. 8-12 1993. Paper 4/00376D Received January 20 1994 Accepted May 24 1994

ISSN:0267-9477    

DOI:10.1039/JA9940901079

出版商:RSC    

年代:1994

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 1085 Aboal-Somoza Manuel 469 Absalan G. 45 Adams F. 151 Akman Suleyman 333 Aller A. J. 871 Alves Luis C. 399 Amarasiriwardena Dula 199 Anderson David R. 67 Anderson S. E. 263 Anghel Sorin D. 635 Anzano Jesus M. 125 Argentine Mark D. 199 Arnold J. T. 263 Arriagada Lorna 93 Arruda Marco A. Z. 657 Avila Akie K. 543 Baaske Bernd 867 Back M. H. 45 Barciela-Alonso Carmen 469 Barinaga Charles J. 1053 Barnes Ramon M. 199 981 Barrios Carlos 535 Barshick Christopher M. 83 Barth Peter 773 Baxter Douglas C. 297 Baxter Malcolm J. 727 Bayne Charles K. 83 Beauchemin Diane 509 Becerra Jose 535 Begley Ian S. 171 Belarra Miguel A. 125 BeneS Petr 303 Bermejo-Barrera Adela 469 Bermejo-Barrera Pilar 469 483 Bernasconi G. 151 Berndt H.861 Berndt Harald 39 193 Betti Maria 385 Bettinelli Maurizio 805 Biffi Claudio 443 Blanco Gonzalez E. 281 Bloxham Martin J. 935 Boge Edward M. 369 Botto Robert I. 905 Branch Simon 33 Brandt R. 1063 Brenner I. B. 737 Briand Alain 17 Broekaert Jose A. C. 1015 1063 Brown Nicole V. 363 Bruhn Carlos G. 535 Bruno Sergio N. F. 341 BudiE Bojan 53 Burakov V. S. 307 Byrne John P. 913 Cabon J. Y. 477 Cai Xiangjun J. 697 Cairns Robert O. 881 Camara Carmen 291 Campbell Michael 187 Campos Reinaldo C. 341 Carey Jeffrey M. 975 Carrion Nereida 205 217 Caruso Joseph A. 145 957 975 Castillo Juan R. 125 311 Cervera Maria Luisa 651 Chakrabarti Chuni L. 45 913 Chartier FredCric 17 Cheam Venghout 315 Chirinos Jos$ 237 483 483 919 CUMULATIVE AUTHOR INDEX JANUARY-SEPTEMBER 1994 Cimadevilla Enrique Alvarez- Cabal 117 Cleland Sandra L.975 Cobo I. G. 223 Coedo A. G. 223 Conver T. S. 899 Cooper 111 C. B. 263 Cordos Emil A. 635 Cornejo Silva G. 93 Cornelis Rita 945 Crews Helen M. 615 727 Cserfalvi Tamas 345 Cujes Ksenija 285 Curtius Adilson J. 341 543 Dadfarnia Shayessteh 7 Dahl Kari 1 Dams Richard 23 177 187 815 1075 Dean John R. 615 De Kimpe Jurgen 945 de la Guardia Miguel 651 Dennis John 727 Denoyer Eric R. 927 Deruaz D. 61 Desrosiers Roland 3 15 Doner Guleren 333 Dorado M. T. 223 Du Xiaoguang 629 Duan Yixiang 629 Durrant Steven F. 199 Ebdon Les 33 611 615 939 Elgersma Jaap W. 619 Eljuri Elias 205 Elmahadi H. A. M. 547 Emteborg Hikan 297 Epler Katherine S. 79 Evans E. Hywel 939 Evans R. Douglas 985 Fadda Sandro 519 Fariiias Juan C. 841 Fecher Peter A. 1021 Feinendegen Ludwig E.791 Feldmann Ingo 1007 Fell Gordon S. 457 Feng Xinbang 823 Fernandez de la Campa M. R. Fernandez Alberto 205 2 17 Fischer Johann L. 623 Fischer W. 257 375 Fisher Andrew S. 611 Florian K. 257 Fonesca Rodney W. 167 Foster Robert D. 273 Foulkes Michael E. 615 Frentiu Tiberiu 635 Gallego Mercedes 657 663 691 Geertsen Christian 17 Ghazy Shaban E. 857 Giessmann Ulrich 1007 Gilmutdinov Albert Kh. 643 Golloch Alfred 867 971 Goltz Douglas M. 919 GomiSEek Sergej 285 Gonzalez Urcesino 535 Goode Scott R. 73 965 Goossens Jan 177 187 Gower Stephen A. 363 369 Gras Nuri T. 535 Greb Ulrich 1075 Greenfield S. 565 Greenway Gillian M. 547 23 1 Gregoire D. Conrad 393 605 Griffin Steven T. 697 Grohs James 927 Guqer Seref 797 Hadgu Negassi 297 Harnly James M.419 Harrison W. W. 991 1039 Hatterer Andre 525 Hauptkorn Susanne 463 Heitmann U. 437 Hernandez Cordoba Manuel Hese A. 437 Hiernaut Tania 385 Hill Steve J. 935 Hinds Michael W. 451 HlavaEek I. 245 251 HlavaCkova I. 245 251 Hoffmann Erwin 685 Holcombe James A. 167 415 Hollenbach Mark 927 Horlick Gary 593 823 833 Houk R. S. 399 Hoult Gavin 7 Howe Alan M. 273 Hu Yanping 213 701 Huang Benli 779 Huang Zhuoer 11 Hudnik Vida 53 Hughes Dianne M. 913 Hutton J. C. 45 Hutton Robert C. 385 881 Imai Shoji 493 759 765 Isaevich A. V. 307 Ito Tetsumasa 1001 Itriago Ana 205 Jackson Jason G. 167 Jakubowski Norbert 193 1007 Janssens K. 151 Jaramillo Victor H. 535 Jepkens Brigitte 193 Jin Qinhan 629 851 Jones Delwyn G. 369 Jung Gerhard 1075 Kabil Mohamed A. 857 Kantor Tibor 707 Karagozler A.Ersin 797 Katskov Dmitry A. 321 431 Kimber Graham M. 267 Kirschner Stefan 971 Kloner A. 737 Kmetov Veselin 443 Koch Lothar 385 Kogan Valentina V. 451 Koirtyohann S. Roy 997 Kolihova Dana 303 Koppenaal David W. 1053 Koropchak J. A. 899 Kratzer Karel 303 Krieger Brian L. 267 Krivan Viliam 463 773 Kroft Marilyn 927 Krug F. J. 861 Krushevska Antoaneta P. 199 Kubova Jana 241 Kujirai Osami 751 Kumamaru Takahiro 89 Kurfurst U. 531 Laborda Francisco 727 Lacour Jean-Luc 17 913 919 553 1015 981 Lamoureux Marc M. 919 Laser Bernd 1075 Lazik C. 45 Le Bihan A. 477 Lechner Josef 3 15 Lee Julian 393 Leis F. 1063 Li Yongquan 679 Li Zhikun 679 Liang Yan Zhong 669 Lile E. S. 263 Liu X. R. 833 Lopez Garcia Ignacio 553 Lopez Jose C. 651 Lopez-Gonzalvez M. Angeles Lord 111 Charles J.599 Ludke Christian 685 Luecke Werner 105 Ma Yizai 679 Mamich Stephen 927 Manickum Colin K. 227 Manninen Pentti K. G. 209 Manzoori Jamshid L. 337 Marais Pieter J. J. G. 321 431 Marchante Gayon Juan Marcus R. Kenneth 45 1029 Marin Sergio R. 93 Martinez-Garbayo Maria Paz Martinsen Ivar 1 Massey Robert C. 61 5 Mauchien Patrick 17 McAllister Trevor 427 McCrindle Robert I. 321 431 McLeod Cameron W. 67 751 Mendes Paulo C. S. 663 Mermet Jean-Michel 17 61 Mezei Pal 345 Michel Robert G. 501 Milagros Gomez M. 291 Miller-Ihli Nancy J. 605 Milton Dafydd M.P. 385 Minnich Michael G. 399 Misakov P. Ya. 307 Mixon Paul D. 697 Moenke-Blankenburg Lieselotte 1059 Moens Luc 177 187 815 1075 Mohl Carola 791 Montoro Rosa 651 Moreda-Piiieiro A. 483 Moreda-Piiieiro J.483 Moreno Rodrigo 841 Mori Toshio 159 Mostafa Mohamed A. 857 Moulton Gary P. 419 Mousty Francis 719 Murillo Miguel 205 217 237 Nagengast Anton 1021 Nagulin K. Yu. 643 Nakahara Taketoshi 159 Nakamura Yoshisuke 75 1 Naoumidis A. 375 Naumenkov P. A. 307 Nevoral Vladislav 241 Ni Zhe-ming 669 Nickel H. 257 375 NiedergesaB Rainer 1071 29 1 713 Manuel 117 1045 125 217 8411086 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1994 VOL. 9 Nobrega J. A. 861 Nolte Joachim 1059 O’Haver Thomas C. 79 419 Ohorodnik S. K. 991 Okamoto Yasuaki 89 Okamoto Yukio 745 Okochi Haruno 751 Olson Lisa K. 975 O’Neill Peter 33 Outred Michael 381 Outridge Peter M. 985 Ozdemir Yuksel 797 Palacios M. Antonia 291 Papaspyrou Manfred 791 Patriarca Marina 457 Pauwels J. 531 Payling Richard 363 369 Peachey Russell M.267 Pepelnik Rudolf 1071 Perez-Arantegui J. 3 1 1 Perez Parajbn Juan M. 11 1 Petit de Pefia Yaneira 691 Petrucci G. A. 131 Petty John D. 267 Pilger C. 1063 Platzner Isaac 719 Polettini Albert0 L. 719 Pollmann D. 1063 Popescu Adrian 635 Poussel E. 61 Prange Andreas 107 1 Pretorius Warren G. 939 Prudnikov Evgeniy D. 619 Pyrzynska Krystyna 801 Quentmeier Alfred 355 QuerrC G. 31 1 Rademeyer Cor J. 623 Radziuk Bernard 1 Raikov S. N. 307 Raith Angelika 1045 Rasmussen Gert 385 Reed Nicola M. 881 Reija Carmen 651 Reyes Olga 535 Richner Peter 985 Rivoldini Alessandro 519 Robles L. C. 871 Rodriguez Aldo A. 535 Romon-Guesnier Sabine 199 RonEeviC Sanda 99 Rosenberg Rolf J. 713 Rowland Stephen J 939 Rubio J. 151 Rummeli Mark H. 381 Sala Jose V.719 Salbu Brit 1 Saleemi Abdollah 337 Salit Marc L. 997 Salud Seremi 535 Santelli Ricardo E. 663 Sanz-Medel Alfredo 11 1 117 Schaldach Gerhard 39 Scheie Andrew J. 415 Schneider Germar 463 Schoknecht G. 437 Schumann Thomas 1059 Schwarzer Rudolph 431 Schwuger Milan J. 791 Segal I. 737 Sekerka Ivan 315 Selby Mark 267 Sharp Barry L. 171 Sheppard Brenda S. 145 Shick Jr. Charles R. 1045 Shtepan Aleksander M. 321 Silva M. M. 861 !ilva R. B. 861 Siroki Marija 99 Sjostrom Sten 17 Skole Jochen 685 Slowick Jeffrey J. 951 Smit Henri C. 619 231 281 Smith B. W. 131 1039 Smith Clare M. M. 419 Smith David H. 83 Smith Fraser O. 267 Smith Monty R. 1053 Smith Trevor A. 67 SpgvaCkova Vera 303 Steers Edward B. M. 381 Steffan I. 785 Stephens Roger 675 Stevenson C . L. 131 StreSko Vladimir 241 Stuewer Dietmar 193 1007 Sturgeon Ralph E.493 605 Su Evelyn G. 501 Sugawa Kazumitsu 89 Sutton Robert L. 1079 Sy T. 437 Takahashi Katsuyuki 75 1 Telgheder Ursula 867 971 Thomas Christopher L. 73 Thomassen Yngvar 1 Thompson K. Clive 7 Tittarelli Paolo 443 805 Tittes Wolfgang 1007 1015 Tomlinson Medha J. 957 Tolg Giinther 1015 1063 Trincherini Pier R. 719 Tsalev Dimiter L. 405 Tschopel P. 1063 Turak Elvan E. 267 Turk Gregory C. 79,997 Uchida Hiroshi 1001 Uden Peter C. 951 Valcarcel Miguel 657 663 691 ValdCs-Hevia y Temprano Vanhaecke Frank 187 Vanhoe Hans 23 177 187 815 1015 759 765 965 M. C. 231 Varga Imre 707 Veber Marjan 285 Verbeek Alistair A. 227 Verrept Peter 1075 Versieck Jacques 23 ViAas Pilar 553 Vincze L. 151 Vujicic G. 785 Wade Jeffery W. 83 Walden W.O. 1039 Walter Serge 525 Wang Jiansheng 957 Wang Jiazhen 679 Wang Xiaohui 679 Wang Xiaoru 779 Wang Ying 851 Webb C. 263 Weiss Zdengk 351 Wiederin Daniel R. 399 Williams J. C. 697 Williams Jr. J. C. 697 Willie S. N. 759 Winefordner J. D. 131 1039 Worsfold Paul J. 611 935 Wrbbel Katarzyna 117 281 Xiao Grace 509 Yang Chenlong 779 Yang Pengyuan 779 Yang Wenjun 85 1 Yu Lijian 997 Yuan Xianglin 851 Zakharov Yu. A 643 Zander A. T. 263 Zaray Gyula 707 Zhang Hanqi 851 Zhang Zhanxia 213 701 Zheng Jianguo 213 701 Zhu Jim J. 905 Zhuang Zhixia 779 Zilkova Jana 303 Zilliacus Riitta 713

ISSN:0267-9477    

DOI:10.1039/JA9940901085

出版商:RSC    

年代:1994

数据来源: RSC