JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 25 1 Multi-element Analysis of Some High Silicon Content Ferroalloys by Inductively Coupled Plasma Atomic Emission Spectrometry* 1. HlavaCkova and I. HlavaCek Analytika Company Limited U Elekfry 650 194 05 Prague 9 Czech Republic ~ ~~ ~ ~~ Procedures using inductively coupled plasma atomic emission spectrometry (ICP-AES) have been developed for both the determination of matrix and minor and trace elements in some FeSi alloys such as FeSi 45 FeSi 75 FeCrSi FeMnSi FeMnCaSi FeTiSi and FeZrSi. Keywords Inductively coupled plasma atomic emission spectrometry; multi-element analysis; ferroalloys; sample decomposition; silicon; microwave digestion In the steel producing industry ferroalloys with high silicon content are used for de-oxidation and final modification of chemical composition of steels before casting.Sample decomposition and the quantitative sample conver- sion into a soluble form are relatively difficult. The FeSi alloys are usually decomposed by fusing with sodium peroxide or by dissolution with hydrofluoric acid. For FeSi alloys analysis using inductively coupled plasma atomic emission spec- trometry (ICP-AES) the solid samples were dissolved using phosphoric acid. Experimental Instrumentation The analyte elements were determined by means of an ARL 33000 LA sequential emission spectrometer with an inductively coupled argon plasma under compromise instrumental con- ditions i.e. at an observation height of 18 mm and an argon carrier gas flow rate of about 1.2 1 min-'. A Perkin-Elmer 593 flame atomic absorption spectrometer was used for compara- tive analysis by flame atomic absorption spectrometry.A commercial domestic microwave oven Philips M 704 with a timer and variable power settings equivalent to outputs of 210 330 450 and 600 W was used. Sample Preparation It was found that phosphoric acid is suitable for the decom- position of carbides nitrides and silicides contained in steels. Gelatinous silicic acid is converted into a solution with phos- phoric acid which also attacks so-called 'metallic' silicon. It was observed that the dissolution of silicic acid gel is relatively faster (about 15 min) than that of 'metallic' silicon. The process rate can be substantially increased if the 'metallic' silicon particle size is about 10 pm. The conversion of silicon into a soluble form with phosphoric acid is very important in the sample preparation of complex high silicon content ferroalloys. In addition silicon can be determined together with the other analyte elements. The general working procedure with phos- phoric acid for the decomposition of metallic samples can be described as shown in Scheme 1.The systems Fe-Si Cr-Si Mn-Si Ti-Si and Zr-Si represent various silicides occurring in high silicon content ferroalloys. So-called 'metallic' silicon is found in alloys containing more than 50% Si. It is similar in the ferrosilicon alloys see Table 2. The metal silicides are relatively soluble in phosphoric acid. The presence of 'metallic' silicon in some FeSi alloys results in a slowing of sample decomposition because sample particles can be coated with insoluble reaction products if the decompo- sition takes too long.Therefore in such cases the sample decomposition was started with sodium hydroxide solution. In principle the FeSi alloy samples in the fine powder form (the sample particle sizes should be less than 0.05 mm without separating the various size fractions) are digested with phos- phoric acid in polytetrafluoroethylene (PTFE) vessels at a temperature of between 150 and 220 "C without the application of hydrofluoric acid. The decomposition time has to be pro- longed for about 10-15 min after sample dissolution. In some instances it is necessary to carry out sample decomposition with sodium hydroxide solution as in the case of FeSi FeCrSi and FeZrSi alloys. After sample dissolution the ferrosilicon alloy sample solutions are oxidized and stabilized with nitric acid or hydrogen peroxide and immediately diluted with de-ionized water to a final volume of 500ml.The sample dissolution procedures should be strictly adhered to see Table 3. In this way the ferrosilicon alloy samples are quanti- tatively converted into a soluble form. Sample decomposition was also carried out in a microwave oven using the same procedures as in the PTFE beakers under atmospheric press- ure. The decomposition time was reduced from 2-3 h to about 10-15 min for all ferrosilicon alloy samples see Fig. 1 and Table 4. A blank and calibration samples were prepared by the same procedures as for the real ferrosilicon alloy samples using certified and internal ferrosilicon alloy reference materials with 150-200 "C I I t not diluted immediately a gel can arise in about Sample solution (syrup) contains a mixture of H3P04 and H4P207 (so-called 'condensed' or 'strong' phosphoric acid) after evaporation I to white fumes of sulfur frioxide to white fumes of sulfur trioxide again Clear and stable sample solution is obtained if it is diluted with water immediately after cooling Hydrated silicon dioxide Si02*x H20 is usually excluded from sample so- lution after dilution * Presented at the XXVIII Colloquium Spectroscopicurn Internationale (CSI) York UK June 29-July 4 1993.Scheme 1252 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 200 9 h' 100 0 5 10 Time/min Fig. 1 Dependence of temperature (7) in a sample solution on heating time in the microwave oven Philips M704 at power outputs of A 600; B 450; C 330 and D 210 W.Measured solution 50ml of phosphoric acid (85% m/m) in a PTFE beaker known additions of some analyte elements in the concentration range to be considered see Table 5. Results and Discussion With regard to the possible presence of non-metallic compo- nents (especially SiO and A1,03 present in slags) in the ferrosilicon alloy samples it was necessary to analyse contin- gent insoluble residues. The residues were analysed after separ- ation by membrane filtration and fusion with sodium carbonate in a platinum crucible. The aluminium and silicon contents determined in the residues were practically negligible (< 0.02% A1 and <0.05% Si).The analyte elements and the spectral lines used are pre- Table 1 Chemical composition of ferrosilicon alloys (%) Alloy FeSi 45 FeSi 75 FeCrSi FeMnSi FeMnCaSi FeTiSi FeZrSi Fe Si Ca Cr 40 40-48 20 72-78 45-55 28-32 15-30 40-50 15-30 15-35 25-50 Mn Ti Zr 60-70 15-30 20-30 30-40 Table 3 Sample preparation sented in Table 6. The spectral interferences of the matrix and the other accompanying elements on the spectral line intensities of the analyte elements were investigated. The more significant effects were found only for iron on chromium manganese on nickel and zirconium on hafnium i.e. the interfering element at a concentration of 1% gives a background equivalent concentration (the concentrations are related to solid samples) of as follows 1% Fe ...0.0005% Cr 1% Mn ... 0.0002% Ni 1% Zr ... 0.0007% Hf. In practice the interferences were corrected if necessary for the increase or decrease in the spectral background due to differences in the sample matrix composition. Special emphasis was placed on the reliability of high matrix element content determination. For comparison and verifi- cation of accuracy the ferrosilicon alloy samples were also analysed by other analytical methods such as gravimetry for Si and Zr (with cupferron) titrimetry for Cr (with potassium permanganate after oxidation with potassium peroxodisulfate in the presence of silver nitrate) Mn (with potassium permanga- nate after separation with zinc oxide) Ca and Fe (with complexone) and Ti [with iron(II1) chloride]. The high matrix element content determination was also examined by means of internal standardization (Ru) for the purpose of precision enhancement.The contribution of internal standardization however was insignificant. Tables 7-11 give a comparison of the ICP-AES results to those of other analytical methods for the ferrosilicon alloy matrix elements. Similarly the results of ICP-AES and other analytical methods such as gravimetry titrimetry molecular absorption spectrometry (UV/VIS spec- troscopy) neutron activation analysis and flame atomic absorption spectrometry (FAAS) are shown in Tables 12 and 13 for both high and low element contents. The ICP-AES analytical procedures were also verified by means of Czechoslovakian and British ferrosilicon alloy certified and internal reference materials (measured against synthetic calibration samples) see Table 5.The precision of determination of analyte elements characterized by the basic statistical data is presented in Tables 14 and 15 for FeSi 45 Table 2 Specification of silicides occurring in ferrosilicon alloys Ferrosilicon alloy Silicide FeSi FeCrSi CrSi CrSi FeMnSi Mn,Si MnSi Mn,Si3 FeTiSi Ti2% TiSi FeZrSi ZrSi Fe3Si FeSi Fe& 'metallic Si' Alloy FeSi 45 FeSi 75 FeCrSi FeMnSi* FeMnCaSi FeTiSi FeZrSi Sample mass/g Pre-decomposition Dissolution 0.300 - 50 ml H3PO4 conc. 0.100 10 ml10Y0 (m/m) NaOH 10 ml HNO conc. 50 ml H3PO4 conc. 0.300 5 ml 10% (m/m) NaOH 50 ml H3PO4 conc. 0.300 - 50 ml H3PO4 conc. 0.300 50 ml H3P04 conc. 0.200 5 ml 10% (m/m) NaOH 10 ml HNO conc. 50 ml H3P04 conc. Oxidation and stabilization 10 ml HNO conc.10 ml HNO conc. 2mlH,02(1+1)or 10 ml HNO conc. +0.2 g NaNO 5ml H,02 (1+1) * Sample preparation the same for FeMnSi and FeMnCaSi.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 253 Table 4 Comparison of the decomposition times for some alloy types Solution Sample type temperaturePC FeCrSi FeMnSi 200-230 FeTiSi A1 150-170 FeTi Ti-alloys 150-170 Ni-steel Co-alloy Real 170-200 ‘Hard’ FeCr (8%C) 200-230 A1 alloys (< 10% Si) 170-200 Decomposition time/min Electrical Microwave heating-plate oven* 120-180 12-15 60- 120 10-12 45-60 10-12 120 10 120-1 80 12-15 60-90 10 * Power output of the microwave oven was 330 W Table 5 Specification of used reference materials Table 7 Comparison of results for silicon content (%) in real and CRM samples of FeSi 75 Si Si ICP-AES ICP-AES Si (Si Ru ratio)? (without Ru) Sample gravimetry* (internal standard) (directly)* C 5619 74.70 74.45 74.60 C 5620 75.20 76.15 75.15 C 5621 76.40 75.80 75.95 77.45 77.35 77.40 C 5624 C 5626 67.70 69.05 69.10 C 5629 76.00 76.35 76.30 C 5630 74.45 76.25 75.90 C 5631 72.85 73.10 72.95 4-1-01 76.98$ 76.80 77.10 BCS-CRM 305 76.00$ 76.10 75.80 CSAN-CRM ~ ~~ Ferrosilicon alloy Reference material FeSi 45 ~SAN-CRM* 4-1-02 FeSi 75 FeCrSi GSAN-CRM 4-5-03 FeMnSi CSAN-CRM 4-5-02 TSAN-CRM 4-1-01 BCS-CRMt 305 FeTiSi Internal RM FeZrSi Internal RM 1st I * Correlation coefficient for gravimetric and ICP-AES (Si measured t Correlation coefficient for gravimetric and ICP-AES (Si Ru ratio) 1 Certified value.directly without Ru) methods was found to be 0.980.methods was found to be 0.967. * CSAN Czechoslovakian Analytical Normal. t BCS British Chemical Standard. Table 6 Analyte elements and analytical spectral lines Table9 Comparison of results for manganese content (YO) in real samples of FeMnSi; correlation coefficient for determination of manga- nese (titrimetry and TCP-AES) was found to be 0.982 Element and line A1 I Ca I Cr I c u I Fe I1 Hf I1 Mn I1 Ni I1 Ru* I Si I Ti I1 Zr TI Wavelength/nm 394.401 422.673 360.53 3 324.754 259.940 264.14 1 257.610 231.604 349.894 251.61 1 323.452 343.823 Secondary slit-width/pm 75 75 75 75 75 75 75 75 75 75 75 75 Sample C 5611 C 5614 C 5617 C 5658 C 5693 C 5694 C 5696 C 6004 C 6005 C 6010 C 6020 Mn titrimetry 73.55 73.20 67.65 68.50 75.65 74.10 73.80 72.80 69.35 68.35 73.50 Mn 73.10 73.65 68.10 68.65 75.80 74.45 73.25 72.75 70.65 69.10 72.90 ICP-AES * Measured on reference photomultiplier (internal standardization). Table8 Comparison of results for silicon and zirconium contents (YO) in real samples of FeZrSi; correlation coefficient for determination of silicon (gravimetry and ICP-AES) was found to be 0.989 and for determination of zirconium (gravimetry and ICP-AES after decomposition of FeZrSi samples with H,PO,) was found to be 0.989 Sample C 5566 c 5574 C 5576 C 5589 C 5591 C 5603 C 5604 C 5605 C 5636 1st I1 F 39 Si gravimetry 47.90 52.40 48.10 5 1 .OO 55.00 56.15 54.50 52.25 43.40 51.20 43.90 Si 48.60 53.60 48.30 52.30 54.60 56.10 54.90 52.50 44.3 5 52.50 43.70 ICP-AES Zr gravimetry 29.15 29.50 37.70 33.45 32.5 5 35.45 36.40 39.55 34.90 35.30 23.70t Zr ICP-AES (H3PO4) 29.10 29.00 37.30 32.70 32.00 35.70 36.20 39.80 34.90 35.60 25.60 Zr ICP-AES (HF)* 29.20 29.45 37.15 33.00 35.20 36.30 39.25 - * Determination of zirconium was performed after decomposition of FeZrSi samples with HF.Determination of zirconium was performed by molecular absorption spectrophotometry with Xylenol Orange after hydroxide separation.254 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 Table 10 Comparison of results for chromium content (YO) in real samples of FeCrSi; correlation coefficient for determination of chromium (titrimetry and ICP-AES) was found to be 0.925 Sample Cr titrimetry Cr ICP-AES Sample Cr titrimetry Cr ICP-AES Sample Cr titrimetry Cr ICP-AES C 5590 C 5609 C 5622 C 5623 C 5627 C 5628 C 5634 C 5635 3 1.90 3 1.90 30.75 31.55 29.20 31.20 29.00 29.60 32.20 32.50 3 1 .OO 31.15 29.15 3 1.50 28.95 29.70 C 5642 C 5644 C 5648 C 5653 C 5654 C 5659 C 5664 C 5666 29.60 29.45 28.75 31.80 27.50 31.40 29.85 30.65 28.90 28.60 28.25 30.70 27.05 30.45 29.95 30.45 C 5667 C 5669 C 5670 C 5671 C 5672 C 5675 C 5676 30.30 30.30 30.45 30.00 30.45 30.75 28.55 30.80 30.20 29.70 29.95 30.30 31.30 28.70 Table 11 Comparison of results for manganese silicon and calcium contents (YO) in real samples of FeMnCaSi Mn Mn Si Si Ca Ca Sample titrimetry ICP-AES gravirnetry ICP-AES titrimetry ICP-AES C 5606 13.70 13.60 52.35 51.90 32.05 31.95 C 5660 23.75 24.00 48.95 49.20 19.40 18.60 C 5681 24.50 24.85 44.90 45.30 18.75 18.05 C 6000 30.70 30.50 41.35 41 .OO 16.85 16.70 Table 12 Comparison of results (%) for real and CRM samples of FeCrSi; correlation coefficients for determination of chromium and silicon were found to be 0.960 and 0.945 respectively Sample C 5562 C 5563 C 5564 C 5583 C 5587 C 5590 C 5609 C 5622 C 5623 CSAN-CRM 4-5-03 Method FAAS others FAAS others FAAS others FAAS others FAAS others FAAS others FAAS others FAAS others FAAS others Certificate value FAAS others ICP-AES ICP-AES ICP-AES ICP-AES ICP-AES ICP-AES ICP-AES TCP-AES ICP-AES TCP-AES A1 0.60 0.64 0.80 0.84 1.05 1.02 - 1.04 1 .00 0.76 1.02 1 .oo - 0.99 0.98 Cr 3 1.30* 31.35 30.30* 29.75 32.75* 32.60 29.15* 29.35 29.60* 29.95 31.90* 32.20 31.90* 32.50 30.75* 3 1 .OO 31.55* 31.15 32.62 32.80* 32.70 c u 0.02 0.025 0.03 0.025 0.02 0.02 0.03 0.03 0.03 0.02 0.015 0.03 0.03 0.025 0.03 0.025 0.05 0.045 0.015 0.02 Mn 0.14 0.13 0.13 0.13 0.13 0.13 0.20 0.18 0.19 0.18 0.10 0.11 0.28 0.28 0.19 0.26 0.25 - 0.08 0.085 Ni 0.175 0.18 0.17 0.17 0.17 0.17 0.17 0.175 0.16 0.16 0.16 0.15 0.17 0.17 0.16 0.17 0.18 0.18 0.15 0.175 - Si 49.35t 49.80 5 1.20t 50.10 50.15t 49.70 47.501- 48.45 49.609 50.20 51.00t 50.80 48.751- 48.80 53.251- 53.05 49.157 49.20 51.60 5 1.40 - Ti 0.0951 0.11 0.081 0.10 0.081 0.085 0.091 0.08 0.091 0.10 0.151 0.13 - 0.14 0.13 0.06$ 0.085 - - * Titrimetry.t Gravimetry. 1 Photometry (with H,O,). Table 13 Comparison of results (YO) for real ferrosilicon alloy samples; sample decomposition for ICP-AES was performed by means of the microwave oven Sample FeSi 75 Fe M n Si/ A FeMnSi/B FeTiSi/A FeTiSi/B FeZrSi/A FeZrSi/B FeZrSi/C Method FAAS others FAAS others FAAS others FAAS others FAAS others FAAS others FAAS others FAAS others ICP-AES ICP-AES ICP-AES ICP-AES ICP-AES TCP-AES ICP-AES JCP-AES A1 2.00 2.06 0.01 < 0.03 0.02 < 0.03 14.35 14.45 7.00 7.20 1.68 1.65 0.68 0.63 0.70 0.74 Cr 0.13 0.12 - 0.05 0.05 0.335 0.32 0.16 0.19 0.40 0.35 0.46 0.51 0.42 0.38 c u 0.06 0.06 0.025 0.03 0.03 0.035 0.23 0.24 0.03 0.03 0.20 0.19 0.20 0.20 0.12 0.12 Fe - 11.os-f 12.05 7.40t 7.301 - 10.657 10.60 - Hf* Mn 0.12 0.13 - 67.80 - 73.50t - 73.55 - 0.95 - 0.96 0.54 - 0.53 0.44$ 0.24 0.46 0.22 0.53$ 0.24 0.51 0.23 0.22 - 0.27 - - - 67.301- - - Ni 0.04 0.04 0.04 0.06 0.07 0.09 0.04 0.05 0.025 0.02 0.03 0.02 0.03 0.04 0.04 0.05 Si 78.3% 78.25 15.90s 15.95 15.65s 16.05 25.8% 26.20 34.30 42.45s 42.40 52.4@ 53.60 51.00s 52.30 33.75s Ti 0.077 0.11 0.1q 0.22 0.2q 0.22 22.80t 23.20 19.907 19.40 1.2q 1.25 1.2017 1.31 0.2017 0.21 Zr - 33.854 34.20 29.5% 29.00 32.70 33.45s * Determination of hafnium was performed after decomposition of FeZrSi samples t Titrimetry. $ Neutron activation analysis.7 Photometry (with H,O,). with HF. 8 Gravimetry.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 255 Table 14 Precision of determination of analyte elements (YO) in FeSi CRM samples Sample CSAN-CRM 4- 1-02 FeSi 45 CSAN-CRM 4-1-01 FeSi 75 BSC-CRM 305 FeSi 75 Parameter* Certificate value n Average SD RSD Certificate value n Average SD RSD Certificate value n Average SD RSD A1 0.49 0.435 0.047 11 10.9 19 1.86 1.905 0.050 2.6 1.25 6 1.27 0.088 6.9 Cr 0.32 0.325 0.007 2.1 0.16 0.155 0.012 7.4 7 0.085 0.004 4.6 12 19 - c u - 15 0.108 0.004 3.5 - 14 0.038 0.004 10.5 8 0.103 0.004 3.7 0.107 Fe - 16 56.91 0.665 1.2 - 18 20.74 0.595 2.9 2 1.4t 6 21.91 0.360 1.6 Mn 0.32 0.325 0.007 2.0 0.11 0.105 0.009 8.1 14 19 - 7 0.135 0.00 1 1 .o Ni - 14 0.045 0.003 6.1 - 18 0.035 0.017 53.5 7 0.075 0.002 2.0 - Si 41.05 12 41.21 0.289 0.7 76.98 15 76.81 0.375 0.5 76.0 6 76.01 0.127 0.2 Ti - 13 0.05 0.006 13 13 - 0.115 0.022 19.4 - 7 0.115 0.002 1.5 * n =No.of analyses; SD = standard deviation; RSD =relative standard deviation. f Informative value not certified. Table 15 Precision of determination of analyte elements (Yo) in CRM samples of FeCrSi and FeMnSi and real samples of FeTiSi and FeZrSi Sample CSAN-CRM 4-5-03 FeCrSi CSAN-CRM 4-5-02 FeMnSi FeTiSi FeZrSi Parameter Certificate value n Average SD RSD Certificate value n Average SD RSD n Average SD RSD n Average SD RSD A1 - 19 0.98 0.015 1.5 - 13 < 0.03 - - 6 14.22 0.18 1.3 7 1.57 0.045 2.9 Cr 32.62 18 32.68 0.327 1 .o - 15 0.3 1 0.049 15.8 - - - - 7 0.525 0.010 1.9 c u - 18 0.019 0.002 9.5 - 17 0.020 0.00 1 6.2 - - - - 7 0.126 0.004 3.2 Fe - 18 14.95 0.109 0.7 - 15 3.89 0.106 2.7 6 35.47 0.29 0.8 Zr 8 34.22 0.28 0.8 Mn - 18 0.085 0.007 7.8 74.08 15 73.98 0.384 0.5 - - - - 7 0.15 0.004 2.7 Ni - 19 0.175 0.025 14.5 16 - 0.045 0.008 17.0 - - - - 7 < 0.02 - - Si 5 1.60 17 51.39 0.455 0.9 19.95 18 19.95 0.286 1.4 6 26.23 0.21 0.8 8 42.39 0.22 0.5 Ti - 17 0.085 0.014 15.8 16 - 0.295 0.028 9.6 6 23.19 0.13 0.6 7 1.79 0.03 1.7 (containing 45% Si) FeSi 75 (containing 75% Si) FeCrSi FeMnSi FeTiSi and FeZrSi.The limits of determination found for ferrosilicon FeSi 75 and the other ferrosiliconalloys are reported in Tables 16 and 17 respectively. Conclusion The ICP-AES procedures can be applied effectively to the multi-element analysis of some ferrosilicon alloys. The pro- cedures proposed are reliable and relatively simple. Silicon can be determined together with the other analyte elements. The application of hydrofluoric acid is not required and therefore Table 16 Limits of determination (defined as ten times the standard deviation of the background noise) for FeSi 75 Limit of determination Element % ng ml-' A1 0.10 200 Cr 0.05 100 c u 0.02 40 Mn 0.02 40 Ni 0.02 40 Ti 0.02 40 Table 17 Limits of determination (defined as ten times the standard deviation of the background noise) for FeSi 45 FeCrSi FeMnSi FeTiSi and FeZrSi Limit of determination Element YO ngml-' A1 0.03 180 Cr 0.02 120 c u 0.01 60 Mn 0.01 60 Ni 0.02 120 Ti 0.01 60 a quartz plasma torch and a glass nebulizer can be used. If a microwave oven is used for the sample decomposition the digestion time is substantially reduced. The described ICP-AES procedures have been used for the multi-element analysis of high silicon content ferroalloys in everyday laboratory practice for the past 5 years. Paper 3/03932C Received July 7 1993 Accepted October 12 1993