184 Analyst, March, 1974, Vol. 99, $9. 184-189 Determination of Trace Amounts of Lead in Steel and Cast Iron by Atomic-absorption Spectrometry with the Use of Carbon Furnace Atomisation BY F. SHAW AND J. M. OTTAWAY (Department of Pure and Applied Chemistry, University of Strathclyde, Cathedral Street, Glasgow, G1 1XL) A method is described for the determination of 1 to 150 p.p.m. of lead in steel and cast iron involving the use of atomic-absorption spectrometry and carbon furnace atomisation. Samples are dissolved in nitric or perchloric acid and analysed directly without pre-concentration. THE determination of levels of lead greater than 0.01 per cent. in steels can be performed satisfactorily by atomic-absorption spectrometry using flame atomisation.192 However, at lower levels the method lacks sensitivity and a Study Group set up by the Chemical Analysis Committee of BISRA preferred a method based on solvent extraction and spectrophoto- metric determination with dithizone, and this method is now adopted as the British Standard Method (BS 1121) .3 Various atomic-absorption procedures have been developed for the determination of low levels of lead in steel and these procedures have been reviewed by Scholes2 and by Hofton and H ~ b b a r d .~ In all instances, complex solvent extraction pro- cedures have been found necessary and the time of analysis can be very long. The presence of small amounts of lead can cause hot rupture during the rolling and forging of stainless steels and can cause the formation of a graphite structure known as Widmanstatten graphite, which may lead to catastrophic failure of cast irons. There is therefore a need for a rapid method for the determination of lead in steels and cast irons in the range 0-0001 to 0.01 per cent.Recently, colleagues at the University of Strathclyde5 have developed a rapid method for determination of lead at these levels based on anodic stripping voltammetry. The great sensitivity of this technique makes possible the direct analysis of solutions of steels with interference only from high concentrations of copper and molybdenum. Owing to the small sample size required, carbon furnace and carbon filament atomisation provide a considerable improvement in absolute sensitivity over flame atomisation for all elements6 However, Fernandez and Manning' also demonstrated an improvement in the limit of detection in terms of concentration by a factor of 50 for the determination of lead by carbon furnace atomisation compared with flame atomisation.It seemed possible, therefore, that this technique might also offer a rapid method for the determination of lead in steel and cast iron, provided that the problems of the matrix could be overcome. Interferences have been reported7,S in the determination of lead by this technique but it has been founds that these are serious only when the solutions are prepared in chloride media. Provided that the lead solutions are prepared in oxy-anion media, e.g., nitric acid, there is no interference from large amounts of iron and other metals.s By using these conditions, we have developed a simple and rapid method for the determination of small amounts of lead in steel and cast iron.EXPERIMENTAL REAGENTS- Reagents of the highest available purity were used throughout. Stock lead solution (100 $.p.m. of lead)-Dissolve 0.16 g of analytical-reagent grade lead nitrate in water, transfer the solution to a 1-litre calibrated flask and dilute to the mark with water and sufficient AnalaR nitric acid to make the final solution M in nitric acid. Iron solution (10 000 p.9.m.) for interference studies-Dissolve 0.5 g of BCS 149/3 iron in 10 ml of 40 per cent. nitric acid, transfer the solution into a 50-ml calibrated flask and dilute to the mark with water. 0 SAC and the authors.SHAW AND OTTAWAY 185 APPARATUS- The instrument used for all measurements was a Perkin-Elmer 306 atomic-absorption spectrometer equipped with an HGA-70 heated graphite atomiser and a deuterium arc back- ground corrector and coupled to an Electronik 19 strip-chart recorder.A Perkin-Elmer Intensitron hollow-cathode lamp was used as the source, The design and operation of the HGA-70 has been described in detail el~ewhere.~g~ Samples are atomised in a graphite tube, 5-3 cm long and 1 cm in diameter, under an argon atmosphere. Samples were transferred to the centre of the tube by means of 50 or 20-4 Eppendorf micropipettes. The HGA-70 has variable time and temperature selectors for sequentially drying, charring and atomising the samples and, once set, this sequence of operations proceeds automatically. In a typical lead determination, the sample or standard is introduced into the centre of the tube and is then dried at 100 "C.Samples can then be charred at an intermediate temperature of 490 "C and are finally atomised at 2200 "C. Only the atomic-absorption signal obtained during the atomisation period is recorded, as shown in Fig. 1. Both the time of drying and the volume of sample introduced into the tube are varied according to the sample being analysed. PROCEDURE A: MILD STEELS AND CAST IRONS IN THE RANGE 0.0001 TO 0.0010 PER CENT. OF Preparation of calibration solutions-Dilute 10 ml of stock lead solution (100 p.p.m.) to 1 litre with water. This solution should be freshly prepared every day. Transfer 0, 1.0, 2.0, 3.0, 4.0 and 5.0 ml of this solution into 100-ml PTFE beakers each containing 0.5 g of BCS 149/3 iron and add 10 ml of 40 per cent.nitric acid. When the iron has dissolved, transfer the solutions into 50-ml calibrated flasks and dilute each solution to the mark with water. The solutions contain the equivalent of 0, 0.0002, 0.0004, 0.0006, 0.0008 and 0.0010 per cent. of lead in steel when 0.5 g of steel sample is used to prepare 50 ml of solution. (ii) Preparation of sample solutions-Weigh 0.5 g of sample into a 100-ml PTFE beaker and dissolve it in 10 ml of 40 per cent. nitric acid. Transfer the solution to a 50-ml calibrated flask and dilute to the mark with water. For the analysis of solutions, follow procedure E, below. LEAD- (i) PROCEDURE B: MILD STEELS AND CAST IRONS IN THE RANGE 0-0010 TO 0.010 PER CENT. OF Pre@aration of calibration solutions-Dilute 10 ml of stock lead solution (100 p.p.m,) to 100 ml with water.Transfer 0, 1.0, 2.0, 3-0, 4-0 and 5.0 ml of this solution into 100-ml PTFE beakers, each of which contains 0-5 g of BCS 149/3 iron, and add 10 ml of 40 per cent. nitric acid. When the iron has dissolved, transfer the solutions into 50-ml calibrated flasks and dilute to the mark with water. These solutions contain the equivalent of 0, 0-0020, 0-0040, 0.0060, 0.0080 and 0,0100 per cent. of lead in steel when 0-5 g of steel sample is used to prepare 50 ml of solution. Preparation of sample solutions-Proceed as in procedure A (ii) above. For the analysis of solutions, follow procedure E, below. LEAD- (i) Prepare this solution freshly every day. (ii) PROCEDURE c: MILD STEELS I N THE RANGE 0.0100 TO 0.0150 PER CENT.OF LEAD- (i) Preparation of calibration solutions-Transfer by microburette 0, 0.50, 0-60, 0-70 and 0.80 ml of the stock lead solution (100 p.p.m.) into 100-ml PTFE beakers, each of which contains 0.5 g of BCS 149/3 iron, and add 10 ml of 40 per cent. nitric acid. When the iron has dissolved, transfer the solutions into 50-ml calibrated flasks and dilute to the mark with water. These solutions, when diluted five-fold, contain the equivalent of 0, 0*0100, 0.0120, 0.0140, 0-0160 per cent. of lead in steel when 0.5 g of sample is used to prepare 250 ml of solution. (e) Preparation of sample solutions-Proceed as in procedure A (ii) above. Before injection of the sample, dilute each solution five-fold. For the analysis of solutions, follow procedure E, below.PROCEDURE D : STAINLESS STEELS IN THE RANGE 0-0010 TO 0-0100 PER CENT. OF LEAD- (i) Pre@aration of calibration solutions-Dilute 10 ml of the stock lead solution (100186 SHRW AND OTTAWAY: DETERMINATION OF TRACE [Analyst, Vol. 99 p.p.m.) to 100 in1 with water. Transfer 0, 1-0, 2.0, 3.0, 4.0 and 5-0 ml of this solution into 100-ml PTFE beakers, each of which contains 0-5 g of BCS 149/3 iron, and add 10 ml of 60 per cent. m/m perchloric acid. When the iron has dissolved, transfer the solutions into 50-ml calibrated flasks and dilute to the mark with water. These solutions contain the equivalent of 0, 0.0020, 0.0040, 0.0060, 0.0080 and 0.0100 per cent. of lead when 0-5 g of steel sample is used to prepare 50 rnl of solution. (ii) Prepayation of sample solutiofzs-Weigh 0.5 g of sample into a 100-ml PTFE beaker and dissolve it in 10 ml of 60 per cent.m/m perchloric acid. Transfer the solution into a 50-ml calibrated flask and dilute to the mark with water. For the analysis of solutions, follow procedure E, below. This solution should be freshly prepared each day. PROCEDURE E : OPERATION OF THE INSTRUMENT- The instrument is operated under the following conditions : Procedure A B, c D I A \ Wavelength/nm . . .. . . . . Lamp current/mA . . . . . . . . Spectral band widthlnm . . .. .. Drying temperature/"C . . . . . . Drying time/s . . .. .. .. Charring temperaturelac .' . .. .. Charring time/s . . .. .. .. Atomisation temperature/'C . . .. Atomisation voltsge/V . . . . .. Atomisation time/s .. .. .. .. Volume of sample solution/pl . . .. Scale expansion . . .. .. * . Argon flow-rate/l min-l (at 40 p.s.i.) . . 383.3 8 0.7 100 40 - - 2200 8 10 60 x 3 1.5 283.3 8 0.7 100 30 - - 2200 8 10 20 x l 1.5 283.3 8 0.7 100 40 490 30 2200 8 10 50 x3 1-5 Sequentially inject samples and standards into the graphite tube and record the atomic- Interpolate sample concentrations from a absorption signal during the atomisation step. calibration graph obtained from the standards. RESULTS AND DISCUSSION INTERFERENCES IN THE DETERMINATION OF LEAD- Studies718 on interferences in the determination of lead by atonlic-absorption spectro- metry using carbon furnace atomisation have indicated that elements such as sodium, iron, calcium and aluminium depress the lead signal when solutions are prepared in chloride media.No interference was found, however, when solutions were prepared in chloride-free media by using nitrate salts and nitric acid. This procedure evidently prevents the volatilisation of lead as a molecular chloride. Use of an oxy-anion medium leads to the formation of the relatively involatile oxides and recent evidencelo suggests that these oxides are efficiently reduced by the carbon from the graphite tube, liberat-ing metal atoms directly in the gaseous state. In the determination of lead in steel at levels down to 0.0001 per cent., a ratio of lo6 would exist between iron and lead concentrations. No interference from 10000 p.p.m. of iron was found on the signal of 0.01 p.p.m. of lead in nitrate media under the conditions described in procedure E.It would therefore be possible to analyse solutions of cast iron and steel by direct comparison with standard solutions that contain only lead in the appro- priate solvent. However, a smoke signal is given by the iron matrix after the atomic- absorption signal of lead, and although the effect of this signal is removed by the background corrector, it was decided, as a precautionary measure, to add the appropriate concentration of pure iron (BCS 149/3) to the calibration solutions to match that in the samples. DETERMINATION OF LEAD IN STEELS AND CAST IRONS- steels and cast irons. The above procedures were applied to the determination of lead in a range of standard Undissolved silica and carbon were allowed to settle to the bottomMarch, 19741 AMOUNTS OF LEAD I N STEEL AND CAST IRON 187 of the calibrated flask before withdrawing the appropriate aliquot for analysis, but the solution could be filtered if required.In our view, analysis at the levels described should be made as simple as possible so as to avoid possible sources of Contamination. The results are shown in Table I. Results obtained by anodic stripping voltainmetry5 are also given for comparison. TABLE I DETERMINATION OF LEAD IN STEELS AND CAST IRONS Lead, per cent. Sample D1 (cast iron) . . D2 (cast iron) . . D5 (cast iron) . . D6 (cast iron) . . D7 (cast iron) . . D8 (cast iron) . . D9 (cast iron) . . BCS 330 (mild steel) BCS 326 (mild steel) BCS 328 (mild steel) BCS 334 (stainless steel) BCS 335 (stainless steel) .. .. .. .. ... . .. .. .. .. .. Ccrtificate value* 0~00012 0.0028 0.0004 0.0022 0.0038 0.0018 0.0070 0.003 0-014 0.015 0-0011 0.0015 Results by procedure indicated A: 0.00013, 0.00010 0.00014, 0.00015 0.00015 B: 0.0031, 0.0029 0-0027, 0.0028 0.0029 A: 0.00035, 0*00037 0.00037, 0*00036 0.00036, 0-00033 B: 0.0024, 0.0025 0.0024, 0.0023 0.0024 B: 0.0039, 0.0041 0.0042, 0.0041 0-0040 B: 0.0018, 0.0020 0.0018, 0.0019 0.0017 B: 0.0070, 0.0077 0.0069, 0*0073 0,0075 B: 0.0025, 0.0026 0.0023, 0.0026 0.0028, 0.0026 C: 0.015, 0.014 0.014, 0.015 0-014, 0.015 C: 0.014, 0-015 0.016, 0.015 0-015, 0,015 D: 0.0010, 0.0010 0~0009, 0.0009 0~0010 D: 0*0015, 0.0014 0.0013, 0.0013 0.0013 Anodic stripping5 - 0.0022 0*0004 0.0024 0.0038 0.0017 0.0066 t o 0.008 0.0022 0.013 0.015 0*0010 0.0013 * The cast iron samples were provided by the British Cast Iron Research Association, Blantyre, Nr.Glasgow and the certificate values for these samples are the analytical results supplied by them which were obtained by a spectrophotometric method. We are grateful to Mr. J. Sneddon for providing these samples and results. The results obtained by atomic-absorption spectrometry are averages from three measurements on each solution. Typical results for sample BCS 326 and a 0.00014 per cent. lead standard are shown in Fig. 1. The results suggest that the method would be satisfactory for the determination of lead in steel and cast iron. A larger range of results was obtained for sample D9 and a similar range was obtained by anodic stripping voltammetry, probably indicating that this sample is inhomogeneous.Ten samples can be analysed in approxi- mately 1 hour, most of this time being taken up in dissolution of the sample; a single sample can be analysed in about 15 minutes.188 SHAW AND OTTAWAY: DETERMINATION OF TRACE [Analyst, VOl. 99 Fig. 1. Atomic-absorption signals for solutions prepared as in procedure A for (A) a standard solution containing the equivalent of 0.0001 per cent. of lead, i.e. 0.01 p.p.m. of lead, in 10 000 p.p.m. of iron (signal includes a small lead blank from BCS 149/3); (B) sample D1 (both A and B a t x 3 scale expansion); and (C) a standard solution containing 0.01 p.p.m. of lead and 10 000 p.p.m. of iron a t x 10 scale expansion showing a small residual smoke smoke signal at D The reproducibility of the method was tested in two ways (Table 11). The reproduci- bility of the instrument was tested by carrying out ten readings on the same sample solution.The reproducibility of the method as a whole was then tested at two concentration levels, by carrying out ten complete analyses of two standard steel samples. The results in Table I1 indicate that the main contribution to the standard deviation is from the instrument repro- ducibility, the relative standard deviation of sample D1 being greater than that of sample BCS 330 as a x 3 scale expansion was used for D1 and no scale expansion for BCS 330. At the concentration levels being determined, the precision is considered to be adequate. There appear to be small differences between the results on samples BCS 330 and D1 in Tables I and 11.However, calculations involving all of the results for these samples only increase the rela- tive standard deviations to 4.2 per cent. for BCS 330 and 13.6 per cent. for D1. The detection limit (20) and sensitivity (1 per cent. absorption) for the lowest concentration range, procedure A, were found to be 0.00004 and 0.000008 per cent. of lead, respectively. At this lowest concentration range, a blank was detected due to lead in the pure iron added to the standard solutions, and this blank was subtracted from the readings on the standard solutions. As mentioned above, if samples are atomised after only a drying step, smoke is given out during atomisation but only after the lead signal has been obtained. Without the background corrector, a large background absorption signal is obtained from the smoke butMarch, 19741 AMOUNTS OF LEAD I N STEEL AND CAST IRON 189 use of the background corrector effectively suppresses this signal.The small residual back- ground signal can be seen at x10 scale expansion in Fig. 1. Investigations designed to reduce the smoke signal by varying the charring temperature and time resulted only in loss of lead during the charring step at temperatures that were still ineffective in removing the smoke. For determinations on solutions in nitric acid, a charring step was therefore omitted. With perchloric acid, which was necessary for dissolving stainless steels, an explosive evolu- tion of smoke occurred during atomisation if no charring step was included. A charring temperature of 490 "C effectively removed most of the smoke from these solutions without a significant loss of lead.TABLE I1 REPRODUCIBILITY TESTS IN THE DETERMINATION OF LEAD Instrument Reproducibility of reproducibility, separate determinations, per cent. of lead per cent. of lead BCS 330 0.0027 0.0028 0.0028 0.0027 0.0027 0.0026 0.0027 0.0027 0.0026 0.0026 Mean . . . . .. .. . . 0.0027 Certificate value . . .. .. . . 0.003 Standard deviation . . .. .. 0-000075 Relative standard deviation, per cent. 2.8 95 per cent. confidence limits, per cent. of lead . . .. . . .. .. - BCS 330 0.0025 0.0026 0.0024 0.0025 0.0026 0.0026 0.0025 0.0026 0.0025 0.0024 0.0025 0.003 0.000082 3.3 f0~00019 D1 0. 000 14 0.000 14 0~00019 0.00014 0~00012 0.00014 0.00015 0.000 16 0.000 13 0*00014 0.00015 0~00012 0~00002 13-3 f 0.000045 our knowledge , no previous publications have described the application of carbon atomisation to the determination of trace elements in iron and steel.The determination To furnace of chromium in steel by use of a tantalum filament has been reported,ll but only at levels of chromium above 0-3 per cent. In the carbon furnace, lead appears to be released readily from the iron matrix and it would seem possible that other volatile elements could be determined by this technique. With less volatile elements, background correction would be essential and the release of the element from the iron matrix might be more difficult. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. REFERENCES Elwell, W. T., and Gidley, J. A. F., Analytica Chim. Ada, 1961, 24, 71. Scholes, P. H., AnaZyst, 1968, 93, 197. Postlethwaite, R. T., Kidman, L., Bagshawe, B., Bills, K. M., Harrison, T. S., and Watt Smith, J. A., J . Iron Steel Inst., 1970, 500. Hofton, M. E., and Hubbard, D. P., Analytica Chim. Acta, 1970, 52, 425. Metters, B., and Cooksey, B. G., Analyst, in the press. Kirkbright, G. F., Ibid., 1971, 96, 609. Fernandez, F. J., and Manning, D. C., Atom. Absorption Newsl., 1971, 10, 65. Shaw, F., and Ottaway, J. M., to be published. Manning, D. C., and Fernandez, F., Atom. Absorption Newsl., 1970, 9, 65. Campbell, W. C., and Ottaway, J. M., submitted for publication. Maruta, T., and Takeuchi, T., Analytica Chim. Ada, 1973, 66, 5. Received December 28th, 1973 Accepted January 17th, 1974