Analyst June 1983 Vol. 108 $9. 717-721 Use of an Argon - Nitrogen Inductively Coupled 717 Plasma for the Analysis of Aluminium Alloys Subsequent to Alkali Dissolution* Jose A. C. Broekaert and Franz Leis Institut fur Spektrochemie und angewandte Spektroskopie Postfach 778,0-4600 Dortmund 1 Federal Republic of Germany and Gungor Dingler Grundenstrasse 65 CH-8247 Flurlingen Switzerland The determination of a series of elements (boron copper gallium iron, magnesium silicon vanadium and zinc) in aluminium samples by inductively coupled plasma optical emission spectroscopy is reported. High-purity aluminium as well as various types of aluminium alloys (A1 - Cu A1 - Mg, A1 - Mg - Si A1 - Si etc.) were brought into solution to give an analyte con-centration of 0.125% m / V with an alkali dissolution procedure.The detection limits for the mentioned elements range from 5 to 150 pg g-l. Both trace elements and major constituents can be determined in the types of aluminium alloys mentioned by using the same calibration graphs. Keywords Alkali dissolution ; aluminium analysis ; argon - nitrogen induc-tively coupled plaswaa ; optical emission spectroscopy The analysis of various types of aluminium samples in the solid state with the aid of X-ray spectroscopic methods or conventional spark optical emission spectrometry generally requires extensive calibration. For off-line analyses sample dissolution and subsequent determinations by flame atomic-absorption spectroscopy are appropriate ; however they may suffer from dynamic range limitations and insufficient power of detection for some important elements (e.g.boron silicon and vanadium). Inductively coupled plasma optical emission spectroscopy (ICP-OES) in our view is a worthy alternative that opens the possibility of simultaneous multi-element analyses. Firstly it enables high-purity aluminium as well as a wide variety of aluminium alloys to be brought into solution; and secondly for alloys with a high silicon concentration it provides a considerable gain in time compared with acid dissolution including the use of fluoric acid as described in ref. 2. According to Greenfield et aL3 (and as applied in this work) the use of a high-power argon -nitrogen ICP may be advantageous because large amounts of sodium are introduced into the solutions when using an alkali dissolution procedure.With this type of ICP a high pressure of aerosol gas (up to 7 bar) can be used which lowers nebulisation effects and the risk of nebuliser blocking in the analysis of solutions with high salt contents. The procedure for the alkali dissolution of aluminium samples described here introduces as little analyte dilution as possible covers a wide variety of alloys and keeps the salt concentration at a tolerable level. Sequential multi-element determinations demonstrate the capabilities of the argon - nitrogen ICP for the trace element analysis of aluminium. Aqalyses were performed over a wide concentration range for various types of aluminium alloy standard sample. Its use for the analysis of aluminium was investigated in this work.An alkali dissolution procedure1 was applied for two reasons. Experiment a1 Instrumentation An argon - nitrogen ICP powered by a free running radiofrequency generator was used. The working coil is an integral part of the oscillator circuit and the output power is stabilised (to within 0.2y0) by feed-back from the radiofrequency stray field which is measured in the * Part of a paper presented at Euroanalysis IV Helsinki August 23rd-28th 1981 718 BROEKAERT et al. ARGON - NITROGEN INDUCTIVELY COUPLED Analyst Vol. 108 vicinity of the coil.* The sample solutions #were freely aspirated using a concentric glass nebuliser (Meinhard ASS.),^ mounted in a glass nebulisation chamber according to Scott et aLs A 0.9-m microcomputer controlled Czerny - Turner monochromator with photoelectric measurement equipment was used for the sequential multi-element analyses.Instrumental details and operational parameters are listed in Table I. TABLE I INSTRUMENTATION ICP-Radiofrequency generator . . FS-10 (Linn supplied by Kontron GmbH); frequency 27.12 MHz (free running); and maximum output power 10 k W with power stabilisation Burner . . . . Three gas flow burner3; outer gas flow 251min-' of nitrogen; Nebuliser . . . . Meinhard Ass. Type B 11; with free aspiration of samples; argon Nebulisation chamber . . . . Dual-wall glass spray chamber Illumination . . . . Three-lens system; and observation zone 4 x 4 mm selected at the intermediate image Spectral apparatus . . . . 0.9-m Microcomputer-controlled Czerny - Turner monochromator; grating 90 x 90 mm a = 1/2400 mm; entrance slit width, 17 pm; exit slit width 25 pm; thermostatically controlled at 30 f 0.1 "C; photomultiplier 9789 QB; quartz refractor plate for background measurements in front of the exit slit4; and microprocessor Intel 80/20 and intermediate gas flow 8 1 min-l of argon flow 2 1 min-l; and argon pressure 4 bar Sample Dissolution Procedure An amount (0.5 g) of sample drillings is transferred into a platinum dish and 15 ml of sodium hydroxide solution (16.5% m/V) prepared from pellets (Merck No.6498) are added. For all types of aluminium alloys investigated (high-purity aluminium A1 - Cu A1 - Mg A1 - Mg - Si, A1 - Si) analyte masses of up to 0.5 g can be completely dissolved. Subsequently the mixture is evaporated to dryness 5 x 1 ml of concentrated nitric acid (Suprapur Merck) are added and after the vigorous reaction a further 10 ml are then added.Aluminium hydroxide is precipi-tated but is re-dissolved after adding 50 ml of distilled water and warming for a few minutes. The clear sample solution is transferred into 100-ml flasks made up to the mark and subse-quently diluted 1 + 3 with distilled water and stored in polyethylene bottles. Using this procedure pure aluminium and all investigated aluminium alloys are rapidly dissolved and their solutions remain stable for several weeks. Analyte solutions contain 0.125% m/V alloy and 0.35% m/V sodium and can be nebulised with a Meinhard nebuliser without the risk of clogging. At these analyte concentrations salt depositions at the tip of the ICP burner did not occur.Results and Discussion Analytical Lines and Detection Limits All measurements were made under working conditions giving optimum line to background intensity ratios. As investigated earlier* they are as follows operating power (defined as the product of plate voltage and current of the radiofrequency generator admitting an efficiency of 0.7) 3 kW; observation zone (4 x 4 mm) located at 4-8 mm above the coil; and aerosol gas pressure 3 4 bar (aerosol gas flow 1-2 1 min-l). The latter were determined according to Kaiser and Specker,lo and were calculated using the equat ionll The selected analytical lines and the obtained detection limits are given in Table 11 June 1983 PLASMA FOR THE ANALYSIS OF ALUMINIUM ALLOYS 719 TABLE I1 DETECTION LIMITS A 3-kW argon - nitrogen ICP was used and the analyte concentration was 0.125%.Aluminium solid samples, Element linelnm % mlm B I 249.7 2.4 x 10-4 Cr I1 286.3 . . 5.5 x 10-4 Cu I 324.8 1.7 x 10-3 Fe I1 259.9 . . 8.2 x 10-4 Mg I1 279.6 . . 2.8 x 10-4 Mn I1 257.6 . . 6.1 x 10-4 Si I 251.6 8.9 x 10-3 V I1 292.4 1.1 x 10-3 Zn I 213.8 4.3 x 10-3 Ga I 294.4 1.5 x Aluminium samples in solution/ng ml-l 3 7 21 51 200 3 8 110 14 53 Aqueous solutions9/ ng ml-l 2 17 37 5 1 2 ---40 where cL is the detection limit c is the background equivalent concentration IB is the blank signal Iu the background intensity and ar(IB + I,) is the relative standard deviation for a series of blank samples.In this work the true background intensity I is determined from the signals measured in the vicinity of the analytical line (as this is possible with the quartz refractor plate technique’). Blank contributions which arise from the reagents used could not always be measured owing to a lack of aluminium samples in which the elements to be determined are completely absent. In this instance they are calculated from the intensity signals for samples in which the concentrations of the respective elements are low. The detection limits referring to the solutions are higher than the values in pure aqueous solutions for iron manganese and magnesium. This may relate to ionisation interferences and electron density changes at high alkali concentrations.12 The lower value for copper may be due to the same reasons.The large difference for iron may arise from the blank contribu-tions. Analysis of Pure and Alloyed Aluminium Samples Determinations of iron and silicon in aluminium (Fig. 1) and of copper in aluminium (Fig. 2) show that widely varying concentrations of third elements (e.g. for the samples in Fig. 1, magnesium concentrations vary from 0.0002 to 2% m/m) do not cause matrix effects; and i+! r‘ 0.5 !!? 0 .- c CI al C u 0 0 2000 0 2 000 Intensity arbitrary units Fig. 1. Determination of (a) iron and (b) silicon in various aluminium alloys. Samples were as follows: <0.2%. mlm; and A1 - Mg (515 525) CMg >2% m/m) (Schweizerische Aluminium AG) . Alkali sample dissolution 0.125% m/ V . Argon - nitrogen ICP 3 kW and Fe(I1) 259.9 nm and Si(1) 251.6 nm lines.A1 (113 114 115 132 134 135 141 142 144) c ~ g 1000 2000 3000 Intensity arbitrary units Fig. 2. Determination of copper in alu-minium alloys. Concentrations (%) referr-ing to theholid samples. Samples were as follows A1 - Mg - Si (614 61G) ; A1 - Mg -Si - Ni (636); A1 - Si (414 416); A1 - Si -Cu - Ni (431 433) (Schweizerische Alu-minium AG). Alkali sample dissolution a 0.0250,b m/V and 0 0.125% m / V . -Argon - nitrogen ICP 3 kW; and a Cu(1) 324.8 nm line 720 BROEKAERT et al. ARGON - NITROGEN INDUCTIVELY COUPLED Analyst Vol. 108 calibration graphs are linear over at least two decades of concentrations. In this work, limitations arose from the dynamic range of the measurement system. It is also shown that a gain in the dynamic concentration range can be obtained by varying the amount of analyte.For example for copper (Fig. 2) samples with low copper contents were dissolved at a con-centration of 1.25 g 1-' and high concentration alloyed A1 - Cu samples at 0.25 g 1-1 and, despite the different concentrations of aluminium in the solutions the same calibration graphs could be used. The method was applied to the analysis of a series of aluminium samples including high-purity aluminium A1 - Mg A1 - Mg - Si A1 - Mg - Si - Mn A1 - Si - Cu - Ni and Unialloy types. For the calibration a series of aluminium standard samples (Table 111) was used. For these calibration samples and for a series of samples to be analysed three replicate measurements were made.The determined concentrations the standard deviations of the estimate and the certified values of the analysed samples are given in Table IV. TABLE I11 CALIBRATION SAMPLES USED FOR THE ANALYSIS OF ALUMINIUM BY ICP-OES All samples Schweizerische Aluminium AG; the results are in Boron . . 0.022 . . 0.007 . . 0.015 ------Chromium Copper 0.0005 -- 0.003 0.021 -- 0.003 - 0.02 0.02 -- -0.2 -- -Gallium Iron 0.013 0.16 - 0.24 - 0.40 - 0.54 - 0.72 - 0.005 5 0.035 -- -Mag- Man-nesium ganese - 0.003 5 - 0.0013 - 0.021 - 0.037 1.93 -- -- -0.1 -- -Sample A1 114 A1 132 A1 133 All34 A1 141 A1 142 A1 144 Unialloy iO7/4 Al-Si414 A1 - Si 416 A1 - Si - Cu - Ni 442 A1 - Mg - Si 614 .. A1 - Mg - Si 616 . . A1 - Mg - Si - Mn 636 Al-Mg515 Silicon 0.096 0.15 0.41 0.52 0.65 ---12.2 13.5 7.9 0.05 0.6 1.2 1.4 Vanadium -0.008 - - -0.045 0.036 -Zinc 0.018 0.014 0.079 0.055 ----- -0.006 -TABLE IV ANALYSIS OF ALUMINIUM BY ICP-OES SUBSEQUENT TO ALKALI DISSOLUTION Results in %' m/m. Certified concentration values (Sck-.veIzerische Aluminium AG) are given in parentheses. Mag- Man-Sample Boron Chromium Copper Gallium Iron nesium ganese Silicon Vanadium Zinc A1115 A1135 All42 Unialloy 207/2 . . Unialloy 207/3 . . A1 - Si - Cu - Ni 431 AI-Mg 511 Al-Mg 525 A1 - Mg - Si 611 . . A1 - Mg - Si - Mn 6:3l A1 - Mg - Si - Mn 633 0.0028 0.002 9 f 0.001 1 (0.0028) (0.004) 0.0047 0.021 f 0.010 f 0.001 (0.0045) (0.02) - 0.039 f 0.001 (0.04) f 0.001 4 - -0.003 9 f 0.0001 (0.0038) -0.002 4 -f 0.0008 (0.0025) 0.0060 -f 0.0008 (0.005 6) 0.038 0.496 f 0.001 * 0.051 (0.37) (0.52) - -0.001 8 -f 0.0005 (0.002 3) 0.0051 0.053 f 0.0005 f 0.003 (0.004) (0.05) - 0.078 f 0.004 (0.079) - -.. 0.011 f 0.002 (0.011) -0.306 f 0.015 (0.31) - 0.517 f 0.015 (0.54) - - 1.58 0.06 (1.61) 1.95 rt 0.05 (1.98) - 7.62 f 0.72 (7.9) - 0.031 - . - . ~ ~ f 0.003 (0.026) - 0.187 f 0.0005 10.181 - 0.062 f 0.001 (0.055) - -0.023 0.287 f 0.006 f 0.014 (0.026) (0.28) - -2.78 f 0.05 (2.87) -- 0.033 f 0.002 (0.033) - - - 0.309 f 0.030 (0.3) - 0.520 f 0.030 (0.53) - 0.890 f 0.030 (0.98 J m e 1983 PLASMA FOR THE ANALYSIS OF ALUMINIUM ALLOYS 721 A standard deviation of the estimate [s(cx)] is calculated from the measured intensities and the concentrations according to Nalimov12 : where cx is the mean of m replicates of the unknown sample as calculated from the regression equation cX = BIx which is obtained from the intensities I measured for n standards with concentrations cl.B is the slope of the calibration graph. The term Z1 is the mean of the c values for all standards and s ( 1 ) is the standard deviation over the regression that follows: Ii are the measured intensities and Ii the intensities for the standard samples calculated from equation (3). The experimental values agree reasonably with certified values.The error of the complete analytical procedure relating to a l-s confidence level is lower than 574 provided that the concentrations being determined are two orders of magnitude above the detection limit which is so for most copper iron magnesium and silicon and some manganese and zinc values. The analytical precision reported as shown by the concentrations of the calibration samples used, can be realised in a range of two decades of concentrations. For the analysis of aluminium alloys with silicon concentrations above 5% m/m an analyte concentration of 0.25 g 1-1 was taken. However the same calibration as for samples with lower silicon concentrations but analyte concentrations of 1.25 g l-l could be used. Conclusion The results show that the use of an argon - nitrogen ICP for alkali dissolution permits trace However the method is limited in some instances by A linear dynamic range of more than two orders of magnitude and low element analysis in aluminium samples.blank contributions. matrix effects make the method attractive for the analysis of random samples. The authors thank Mrs. W. Bartusch for contributing to the experimental work. The standard aluminium samples were kindly supplied by Schweizerische Aluminium AG. The work was supported by the Ministerium fur Wissenschaft und Forschung des Landes Nordrhein-Westfalen and by the Bundesministerium fur Forschung und Technologie. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References Wilson L. Anal. Chim. Acta 1968 40 603. Bell G. F.At. Absorpt. Newsl. 1966 5 73. Greenfield S. Jones I. Ll. and Berry C. T. Analyst 1964 89 713. Dorn G. in Barnes R. M. Editor “Developments in Atomic Plasma Spectrochemical Analysis, Proceedings of the International Winter Conference 1980 San Juan Puerto Rico,” Heyden, Philadelphia 1981 p. 369. Meinhard J. E. ICP Inf. Newsl. 1976 2 163. Scott R. H. Fassel V. A, Kniseley R. N. and Nixon D. E. Anal. Chem. 1974. 46 75. Nordmeyer N. 2. Anal. Chem. 1967 225 247. Broekaert J. A. C. Leis F. and Laqua K. in Barnes R. M. Editor “Developments in Atomic Plasma Spectrochemical Analysis Proceedings of the International Winter Conference 1980 San Juan Puerto Rico,” Heyden Philadelphia 1981 p. 84. Aziz A. Broekaert J. A. C. and Leis F. Spectrochim. Ada Part B. 1982 37 369. Kaiser H. and Specker H. 2. Anal. Chem. 1956 149 46. Aziz A. Broekaert J. A. C. and Leis F. Spectrochim. Ada Part B 1981 36 261. Kalnicky D. J . Fassel V. A. and Kniseley R. N. Appl. Spectrosc. 1977 31 369. Nalimov V. V. “The Application of Mathematical Statistics to Chemical Analysis,” Pergamon Received August 18th 1982 Accepted January 14th 1982 Press Oxford 1963