DECEMBER, 1971 THE ANALYST Vol. 96, No. I149 A Coniprehensive Scheme for the Analysis of a Wide Range of Steels by Atomic-absorption Spectrophotometry* BY D. R. THOMERSON AND W. J. PRICE (Pye Unicam Ltd., York Street, Cambridge, CB1 2 P X ) The determination of manganese, nickel, chromium, molybdenum, copper, vanadium, cobalt, titanium, tin, aluminium and lead in steel by a direct a ‘omic-absorption method, which involves a single dissolution based on perchloric acid, is described. The scheme effects considerable savings in time compared with traditional methods, and is of comparable accuracy. The serious depressive interferences caused by iron on the response of some elements are overcome in the nitrous oxide - acetylene flame, and minor effects are corrected for by inclusion of iron in calibration standards.No other inter-elemental interferences were encountered in perchloric acid based solutions in the presence of iron. Silicon and tungsten, which are not retained in solution in this scheme, are determined after a separate dissolution of the sample. Results obtained with British Chemical Standard steels are tabulated, and accuracies discussed. THE problems arising in the determination of chromium and molybdenum have hitherto prevented the development of a general scheme for the analysis of irons and steels by atomic absorption. A recently published methodl in which these problems appear to have been successfully overcome has provided the clue to such a scheme, in which all the elements commonly encountered in steels can be determined by simple direct procedures.The scheme covers the determination of twelve elements, all of which, with the exception of tungsten, are determined on a single sample solution. A l-g sample of steel is dissolved in hydrochloric and nitric acids and the solution is finally evaporated to fumes with perchloric acid. The sample solution is then compared directly with calibration standards prepared by adding aqueous solutiom of the elements to l-g portions of pure iron and dissolving the latter in a manner identical to that used for the samples. In this way manganese, nickel, chromium, molybdenum, copper, vanadium, cobalt, titanium, tin, aluminium and lead were determined in a wide range of steels. An alternative procedure for the determination of tungsten and for molybdenum in the presence of tungsten is necessary.This procedure is based on the use of a mixture of perchloric, phosphoric and sulphuric acids, which retains the tungsten in solution. APPARATUS- All measurements were made with a Unicam SPSOA, Series 2, atomic-absorption spectro- photometer incorporating an SP91 lamp turret accessory and an SP94 nitrous oxide accessory. Air was supplied through an SP93 air compressor and the nitrous oxide and acetylene from cylinders. The instrument was fitted with an inert nebuliser and Unicam hollow-cathode lamps. The absorbance peaks were displayed on an SP22 chart recorder. The only non-standard modification was the fitting of 0-125-mm metal shims to the nitrous oxide burner to increase the jaw width from 0.46 to 0.59 mm.This was necessary to prevent blockage by solutions of steel of concentrations higher than about 0.5 per cent. The fitting of the shims to the nitrous oxide - acetylene burner naturally reduces to some extent the previous operating safety levels. Therefore, it is important to maintain the gas flow-rates at about the recom- mended values. On no account should shims that are thicker than 0.125 mm be used; the maximum burner jaw width is therefore 0-59 mm. No modification to the air - acetylene burner was necessary. * Presented at the Third SAC Conference, 1971, Durham. 0 SAC and the authors. 825826 THOMEKSON AND PRICE : A COMPREHENSIVE SCHEME FOR THE ANALYSIS OF [Analyst, Vol. 96 MANGANESE- Belcher and Kinson2 reported a slight enhancement of manganese absorption in 10 per cent. nitric acid solution and a slight depression in 10 per cent.solutions of hydrochloric, sulphuric and orthophosphoric acids, and in mixtures of phosphoric and sulphuric acids. They also noted depression by 1 per cent. of iron, whereas we found that the presence of 1 per cent. of iron slightly enhanced the signal. They also noted interference by 20 per cent. of chromium in a sulphuric acid - orthophosphoric acid mixture, which was overcome by using a smaller section of the flame, by placing an iris at the front of the hollow-cathode lamp. This interference did not occur with an air - acetylene flame in our perchloric acid medium as we obtained excellent results on a steel containing 4 per cent. of chromium (B.C.S. 341). However, Belcher and Kinson’s results do show that very good accuracy and reproducibility can be achieved with atomic absorption. We found no interference from either aluminium or molybdenum and, among other workers, only Atsuya3 found it necessary to add these elements to the calibration solutions.Ramirez-Munoz and Roth4 reported that better results were obtained with standards prepared from standard steels than with synthetic standards. Hubbard and Monks6 used a mixture of hydrochloric and nitric acids as the solvent and finally evaporated the solution to fumes with perchloric acid. They found that perchloric acid interfered to a lesser extent than any other acid (which confirms our findings) but reported small interferences by large amounts of cobalt, tungsten, chromium, nickel and molybdenum in the air - acetylene flame.All these interferences, together with the small depressive effect these workers reported for iron, can be completely overcome by the addition of 50 per cent. aqueous ethanol solution, which also serves to increase considerably the sensitivity . Of the many other ~ o r k e r s , ~ - l ~ most used a hydrochloric acid - nitric acid solvent for samples and aqueous calibration solutions containing only additions of iron. Results obtained were usually excellent for standard steels, which confirms our view that few difficulties are experienced with manganese in steel. As some pure irons contain a relatively high proportion of manganese this treatment, coupled with the good sensitivity of manganese, can result in a large blank value, e.g., B.C.S.260/2 containing 0.013 per cent. of manganese gave an absorbance reading of 0.06 for a calibration range of 1 to 20 mg 1-l. This signal, which results entirely from the manganese absorption and not from any interference effects in the flame, must be subtracted from each of the calibration absorbance values. The problem is reduced by using pure iron containing less manganese when available, e.g., B.C.S. 260/3 containing 0.002 per cent. of manganese gives an absorbance signal of less than 0.01, which in most instances can be ignored without adversely affecting the results. To assess the reproducibility of the manganese determination and also the long-term stability of the solutions, the four manganese determinations below were repeated with the same solutions on each of ten successive days.A statistical analysis of the results obtained yielded the following figures. Manganese, per cent. Coefficient of variation, B.C.S. No. (certificate value) Standard deviation per cent. 321 0.13 0.0013 1.04 312 0.20 0.0037 1.77 341 0.43 0.0076 1.83 23512 0.89 0-0090 0.96 NICKEL- The determination of nickel in steel with the air - acetylene flame has usually been found to be easy and interference free. Kinson and Belcherls proposed the use of phosphoric acid - sulphuric acid to retain tungsten in solution, as did Knight and Pyzyna.16 The latter observed that a slight depression of the nickel absorption was caused by most acids and it became necessary to incorporate iron in their standards. Other wm-kers8-15JgJ0 concerned with the determination of nickel in steel used a hydro- chloric acid - nitric acid solvent, which acts more rapidly and has the advantage of dissolving all of the copper whereas a phosphoric acid - sulphuric acid mixture may not.Nickel absorption wavelengths are discussed and some workers1°-13 found it preferable to use theDecember, 19711 A WIDE RANGE O F STEELS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 827 more sensitive line at 232.0 nm for levels of nickel below about 1 per cent., but they noted that the calibration graph for the 341.5 nm line is far more linear than that for the 232.0 nm line. We found that nickel was similar to manganese in giving rise to a small blank value, but whereas the manganese blank resulted entirely from manganese impurities in the iron, with the nickel it was caused only by physical interferences in the flame.However, this effect was completely overcome in the calibration without adversely affecting the analytical results. CHROMIUM AND MOLYBDENUM- In a recent paper1 the present authors described how difficulties encountered in the determination of chromium and molybdenum in steel are overcome. Depression of absorption response is minimised by using nitrous oxide - acetylene instead of air - acetylene flames. Variations caused by the difference in oxidation state of molybdenum and chromium between the samples and standards are overcome by evaporating both samples and standards to fumes with perchloric acid. COPPER- Wallace21,22 obtained high results for copper by using the air - acetylene flame and dis- solving the sample in hydrochloric acid; with phosphoric acid - sulphuric acid solvent the values were satisfactory.Kinson and BelcherZ3 found that although iron suppressed the copper signal no other interferences were present. Phosphoric acid - sulphuric acid solvent, coupled with the addition of iron to the calibration solution, gave good analytical results. Since then, as with manganese and nickel, most ~ 0 r k e r ~ ~ ~ ~ - - 1 ~ s ~ ~ J 6 , ~ * have preferred the hydrochloric acid-nitric acid solvent, combined with the use of either standard steels or pure iron added to aqueous copper solution for the calibration solutions. Exceptionally, Atsuya3 extracted copper with isobutyl methyl ketone as the copper diethyldithiocarbamate complex. Good analytical results were obtained with no detectable interferences.We found that the presence of 1 per cent. of iron depressed the copper absorp- tion in perchloric acid by about 10 per cent., but satisfactory results were obtained when iron was incorporated in the calibration solutions. VANADIUM- Capacho-Delgado and Manning25 successfully determined vanadium in steels by a direct method in which a nitrous oxide - acetylene flame and aqueous vanadium calibration solutions were used. They used the phosphoric acid - sulphuric acid mixture for dissolution of the samples and noted that sulphuric acid depressed the vanadium signal, while phosphoric acid enhanced it. They assumed that these two effects cancelled each other out as their published results were in good agreement with certificate values.In perchloric acid solution, we found that the presence of 1 per cent. of iron enhanced the vanadium signal by about 20 per cent., thus necessitating the addition of iron to the calibration solutions. However, we found no evidence that any other inter-elemental inter- ferences affected the response. As with some of the above elements a small blank value (about 0.03 absorbance unit on a range of 10 to 100 mg 1-l) was evident. After aspirating at a non-absorbing wavelength it was shown that this blank resulted entirely from physical interferences by iron in the flame operating on both samples and standards and had no adverse effects on the precision of the determination when disregarded. COBALT- McPherson, Price and Scaife,26 who used a 2-g sample dissolved in hydrochloric acid - nitric acid and diluted the solution to 50m1, reported a sensitivity of 0.001 per cent.of cobalt in steel with an air - acetylene flame. They also used phosphoric acid - sulphuric acid to dissolve a wide range of high and low-alloy steels when higher levels of cobalt were present. In most subsequent publications on cobalt in ~ t e e 1 ~ , ~ ~ , ~ ~ , ~ ~ , ~ ~ , ~ ~ , ~ ~ no interference problems were reported. We found that in the perchloric acid solvent, the cobalt response was enhanced by about 26 per cent. in the presence of 1 per cent. of iron. A moderate blank resulted from the cobalt calibration, which was caused almost completely by physical interferences in the flame.828 THOMERSON AND PRICE: A COMPREHENSIVE SCHEME FOR THE ANALYSIS OF [Analyst, vol.96 Some pure irons, however, contain appreciable amounts of cobalt, so it is necessary to choose calibration material with care. If cobalt-free iron is not readily available it will be necessary to determine how much of the signal is caused by the physical interference (matrix effect) of iron and how much by the cobalt absorption. To differentiate between these effects the blank solution should be aspirated and measured at a non-absorbing wavelength emitted by the cobalt lamp, with identical scale expansion, which will show how much of the signal results from matrix interference, the difference between this signal and the blank signal at 240nm being caused by the absorption of the cobalt in the iron blank.For most determinations, however, by using a pure iron similar in composition to B.C.S. 260/2 (0.009 per cent. of cobalt), satisfactory results are obtained by ignoring the blank, which in this instance is caused almost completely by matrix interference. TITANIUM- Atomic absorption was successfully applied to the determination of titanium in steel only after the advent of the nitrous oxide - acetylene flame. Bowman and Willis28 dissolved samples in hydrochloric acid - nitric acid and evaporated the solutions to fumes with sulphuric acid. They noticed that the interference effect of iron was a function not only of the sulphuric acid concentration, but also of the flame conditions. Calibration solutions had to be made up so that they contained iron, nickel, chromium and cobalt in concentrations similar to those of the samples.Headridge and H ~ b b a r d ~ ~ dissolved the sample in hydrofluoric and nitric acids, finally making up the solution so that it contained 50 per cent. of ethanol (a similar approach was used5 for manganese). This treatment enhanced the titanium response and overcame all inter-elemental interferences except that of iron, which still had to be added to the calibration solutions. Low results were reported by Mostyn and Cunningham30 when using either hydrochloric or sulphuric acid as a solvent. But when using hydrochloric acid - nitric acid mixture and making a single addition of potassium chloride, all the interferences, including that by iron, were overcome. It was therefore not necessary to match the iron content of samples and standards.With perchloric acid solvent we experienced no interference problems for titanium from any of the more usual elements in steel; even iron had no effect. We therefore confirm that unmatched aqueous standards can be used satisfactorily. However, our titanium values given here were obtained with iron present in the standards, as mixed standards were used throughout. TIN- No previous papers dealing with the determination of tin in steel by atomic-absorption spectrophotometry could be found, probably because of the very poor sensitivity of tin, which makes its direct determination at low levels difficult. With a 2 per cent. sample solution in perchloric acid we were able to determine down to about 0.01 per cent. of tin by using a nitrous oxide - acetylene flame and a wavelength of 224.0 nm.A small enhancement from 1 per cent, of iron was apparent, but no other interferences arose. The noise level of the signal, however, was still rather high, even when operating with maximum damping facilities. ALUMINIUM- The sensitivity of aluminium in the air - acetylene flame is so poor that it was not seriously considered for useful application in steel analysis, especially as most steels contain only very small amounts of aluminium. Before the introduction of the nitrous oxide - acetylene flame, Nikolaev31 distilled alu- minium from solid samples of refractory metals into a graphite cell that was heated to a high temperature, which readily permitted the detection of to per cent. of aluminium. By using the nitrous oxide - acetylene flame, Amos and Thomas32 reported good results for aluminium in the range 0.1 to 6 per cent.after adding iron to the calibration solutions. Pricell described the use of hydrochloric acid - nitric acid for the direct determination of acid-soluble aluminium, which necessitated fusing the residue and adding the resulting solution to the main sample solution for the determination of the total aluminium. A similar method was also used by other workers.24~33~34 Clarke and Cookel2 improved the sensitivityDecember, 19711 A WIDE RANGE OF STEELS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 829 of the determination by using an organic solvent to give a 2-fold enhancement of the signal. A 10 or 20-fold concentration of the aluminium was also demonstrated after extracting the aluminium as cupferrate into isobutyl methyl ketone, which solution could be aspirated directly.In perchloric acid medium, we found that the presence of 1 per cent. of iron gave rise to an enhancement of about 15 per cent. of the aluminium absorption, giving adequate sensitivity for the steels analysed. LEAD- As early as 1961, Elwell and G i d l e ~ ~ ~ used the hydrochloric acid - nitric acid dissolution technique to obtain a 2 per cent. sample solution, which was compared with a range of aqueous lead calibration solutions and corrected by the response of one iron-based standard. The absorption line 283.3 nm was used and good analytical values were reported for the range of 0-05 to 0.5 per cent. of lead. Dagnall, West and Young36 removed the iron matrix with isopentyl acetate and then extracted lead, as the iodide, into isobutyl methyl ketone.This technique yielded good results down to 0.001 per cent. of lead in a wide range of steels. Pricell and Clarke and Cooke12 found that a 2 per cent. sample solution in hydrochloric acid - nitric acid, aspirated into the air - propane flame at a scale expansion of five times on an SP90 atomic-absorption spectro- photometer, enabled lead in the range 0.0025 to 0.015 per cent. to be successfully determined in steel and cast iron. In our perchloric acid solution, the presence of 1 per cent. of iron enhanced the lead response by about 35 per cent., which, coupled with the use of the more sensitive 217.0 nm wavelength, enables lead levels down to 0.001 per cent.to be determined directly on a 2 per cent. sample solution. TUNGSTEN- By using a mixture of sulphuric, phosphoric and perchloric acids to effect sample dis- solution and with standard steels for calibration purposes, Knight and PyzynalG reported good agreement with other methods of analysis at about the 1 to 6 per cent. of tungsten level. We found that the effect of adding 1 per cent. of iron to a solution of sodium tungstate (50 p.p.m. of tungsten) was to suppress the tungsten signal by about 50 per cent., whereas in the presence of perchloric acid - phosphoric acid - sulphuric acid (mixture A, Method 2) the signal was enhanced, resulting in an over-all slight suppression in our final solutions. With this direct method it should be possible to determine any level of tungsten down to about 0.1 PROPOSED METHOD REAGENTS- 50ml of hydrochloric acid (sp.gr.1.18) and dilute to 1 litre. 5 N nitric acid and dilute to 1 litre. Stock manganese solution, 1000 mg I-l-Dissolve 1.0000 g of pure manganese Stock nickel solution, 1000 mg I-l-Dissolve 1-0000 g of pure nickel metal in per cent. metal in 40ml of Stock chromium solution, 1000 mg I-l-Dissolve 1.0000 g of pure chromium metal in 30 ml Stock molybdenum solution, 1000 mg kl-Dissolve 1.829 g of analytical-reagent grade Stock copper solution, 1000 mg Z-l-Dissolve 14000 g of pure copper metal in 50 ml of Stock vanadium solution, 1000 mg I-l-Dissolve 2.296 g of analytical-reagent grade am- Stock cobalt solution, 1000 mg I-1-Dissolve 1.0000 g of pure cobalt metal in 50 ml of 6 N Stock titanium solution, 1000 mg I-l-Dissolve 7.394 g of analytical-reagent grade potas- Stock tin solution, 1000 mg 2-l-Dissolve 1.0000 g of pure tin metal in 50 nil of hydro- Add 150 ml of hydrochloric of hydrochloric acid (sp.gr.1.18) and dilute to 1 litre. ammonium molybdate in water and dilute to 1 litre. 5 N nitric acid and dilute to 1 litre. monium vanadate in 20 ml of 100-volume hydrogen peroxide and dilute to 1 litre. nitric acid and dilute to 1 litre. sium titanium oxalate in water and dilute to 1 litre. chloric acid (sp.gr. 1.18) plus 5 ml of nitric acid (sp.gr. 1.42). acid (sp.gr. 1.18) and dilute to 1 litre.830 THOMERSON AND PRICE : A COMPREHENSIVE SCHEME FOR THE ANALYSIS OF [Analyst, vol. 96 Stock aluminium solution, 1000 mg I-1-Dissolve 1.0000 g of pure aluminium metal in 25 ml of hydrochloric acid (sp.gr.1-18) plus a few drops of nitric acid (sp.gr. 1.42) and dilute to 1 litre. Stock lead solution, 1000 mg 1-1-Dissolve 1.0000 g of pure lead metal in 10 ml of 2 N nitric acid and dilute to 1 litre. Stock tungsten solution, 1000 mg 1-1-Dissolve 1.420 g of ammonium tungstate in water and dilute to 1 litre. Dilute solutions, when needed, are prepared by diluting the above concentrated solutions, and should be prepared daily. Stock iron solution, 5 per cent.-Dissolve 5 g of high purity iron (B.C.S. 260/3) in 40 ml of hydrochloric acid (spgr. 1.18) plus 5 ml of nitric acid (sp.gr. 1.42). When the reaction is complete, add 20ml of perchloric acid (sp.gr. 1.54) and evaporate until fumes of perchloric acid just appear.Cool and dilute to 100ml with water. Perchloric acid, sp.gr. 1.54. Hydrochloric acid, sp.gr. 1.18. Nitric acid, sp.gr. 1-42. Perchloric acid - phosphoric acid - sulphuric acid (mixture A)-To 300 ml of water, add 100 ml of perchloric acid (spgr. 1.54), 100 ml of phosphoric acid (sp.gr. 1.75) and 100 ml of sulphuric acid (sp.gr. 1.84). METHOD 1. DETERMINATION OF MANGANESE, NICKEL, CHROMIUM, MOLYBDENUM, COPPER, Preparation of sample solutions-Weigh 1.0000 g of sample into a 250-ml beaker and dissolve it in 10ml of hydrochloric acid (spgr. 1.18) plus 5ml of nitric acid (sp.gr. 1.42). After the initial reaction has subsided, add 10 ml of perchloric acid (sp.gr. 1.54) and evaporate the solution until it is fully oxidised and fumes of perchloric acid appear.(This is achieved when the solution turns red and perchloric acid is seen to reflux on the sides of the beaker. For samples not containing chromium, the solution will not turn red but is merely allowed to fume for 5 minutes as specified below.) Allow to fume for about 5 minutes, cool and dissolve the soluble salts in about 50 ml of water. Filter the solution through a Whatman No. 541 filter-paper, wash well with water, adding the washings to the filtrate, and dilute to 100ml. Preparation of calibration solutions-To each of seven 250-ml beakers transfer 1.0 g of pure iron and suitable volumes of stock solutions (Table I). Use the same dissolution procedure as specified above for the samples. It is imperative that the chromium and molybdenum stock solutions be added before fuming takes place. However, if the chromium and molybdenum are to be omitted from VANADIUM, COBALT, TITANIUM, TIN, ALUMINIUM AND LEAD- TABLE I CALIBRATION SOLUTIONS These are based on a 1.0-g sample (except when otherwise stated) and a final volume of 100ml is used throughout Manganese- Dilute stock solution (200 mg l-l)/rnl Concentration/mg 1-1 .. .. Manganese, per cent. .. .. Concentration/mg 1-1 . . .. Nickel, per cent. . . .. .. Concentration/mg 1-1 . . . . Chromium, per cent. .. . . Concentration/mg 1-1 . . .. NickeJ- Stock solution (200 mg l-l)/ml . . Chromium ( u p to 0-5 per cent.)- Stock solution (500 mg 1-f)/ml . . Chromium (up to 4.0 per cent.)- Stock solution (1000 mg l-l)/ml . . Chromium, per cent. (0-25-g sample) 0 1.0 0 1 0 0.01 0 1.0 0 2 0 0.02 0 1.0 0 5 0 0.05 0 1.0 0 10 0 0.4 2.0 2 0.02 2.5 5 0.05 2.0 0.1 2.0 0.8 10 20 5.0 5 0.05 5.0 0.10 4-0 0.2 4.0 1.6 10 20 40 10.0 10 0.10 10.0 20 0.20 6.0 0.3 6.0 2.4 30 60 15.0 15 0.15 15.0 30 0.30 8.0 0.4 8-0 3.2 40 80 20.0 20 0.20 20.0 40 ._ 0-40 10.0 50 0-5 10.0 4.0 100December, 19711 A WIDE RANGE OF STEELS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 831 TABLE I-continued Chromium (up to 12.0 per cent.) *- Stock solution (1000 mg l-l)/ml .. Chromium, per cent. (0.25-g sample) Dilute stock solution (100 mg l-l)/ml Concentration/mg 1-l . . .. Molybdenum (up to 0.2 per cent.)- Concentration/mg 1-1 . . .. Molybdenum, per cent. . . .. Concentration/mg 1-' . . .. Molybdenum, per cent. . . .. Concentration/mg 1-1 . . .. Molybdenum, per cent.. . .. Concentration/mg 1-l . . .. Copper, per cent. . . .. .. Concentration/mg 1-1 . . .. Vanadium, per cent. .. .. Concentration/mg 1-' . . .. Cobalt, per cent. . . .. .. Concentration/mg 1-' . . .. Titanium, per cent. .. .. Concentration/mg 1-1 . . .. Tin, per cent. .. .. .. Concentration/mg 1-1 . . .. Aluminium, per cent. . . .. Concentration/mg 1-1 . . .. Lead, per cent. . . .. .. Concentration/mg 1-l . . . . Tungsten, per cent. . . .. .. Molybdenum (up to 0.5 per cent.)- Stock solution (500 mg 1-I)In-d . . Molybdenum (up to 1.0 per cent.)- Stock solution (1000 mg l-l)/ml . . Copper- Dilute stock solution (100 mg l-l)/ml Vanadium- Stock solution (1000 mg l-l)/rnl . . Cobalt- Stock solution (500 mg l-l)/ml . . Titanium- Stock solution (1000 mg l-l)/ml . . Tin- Stock solution (500 mg l-I)/rnl .. Aluminium- Stock solution (500 mg l-l)/ml . . Lead- Dilute stock solution (100 mg l-l)/ml Tungstent- Stock solution (1000 mg l-l)/ml . . 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.5 25 1 2-0 2 0.02 1.0 5 0.05 1.0 0.1 1.0 1 0.01 1.0 0.1 1.0 5 0.05 1.0 0.10 1.0 5 0.05 1.0 5 0.05 1.0 1 0.01 5.0 0.5 10 10 10 50 5.0 50 2 5-0 5 0.05 2-0 0.1 2.0 0-2 2.0 2 0-02 2.0 0.2 2-0 0.10 2.5 0.25 2.0 0.10 2.0 0.10 2-0 2 0.02 10 20 20 10 25 10 10 10.0 100 1 7.5 75 3 10.0 10 0.10 4.0 0.2 4.0 0.4 5.0 5 0-05 4.0 0.4 4.0 0.20 6.0 0.50 4.0 0.20 4.0 0.20 5.0 5 0.05 20 40 40 20 50 20 20 20.0 200 2 10.0 100 4 15.0 15 0.15 6.0 0.3 6.0 0.6 10.0 10 30 60 0.10 6.0 0.6 6-0 0.30 60 30 10.0 1.0 6.0 0.30 6-0 0.30 100 30 30 10.0 10 0.10 30.0 300 3 20 200 8 20.0 20 0.20 8.0 0-4 8.0 0.8 15.0 15 40 80 0.15 8.0 0.8 8.0 0.40 80 40 15.0 1.6 8.0 0.40 8.0 0.40 150 40 40 10.0 15 0.15 40.0 400 4 30 300 12 10.0 50 0.5 10.0 1.0 20.0 20 100 0.20 10.0 1.0 10.0 50 100 0.50 20.0 2.0 10.0 50 200 0.50 10.0 50 0.50 20.0 20 0.20 * For this range the burner must be in the fully rotated position.t 50 ml of acid mixture (A) must be added to each of the tungsten calibration solutions (Method 2). the scheme, the stock solutions can be added after the iron has been dissolved for the calibra- tion ranges. Alternatively, the pure iron stock solution can be added to the aqueous standards before dilution (i.e., 20 ml of pure iron stock solution in a final volume of 100 rnl is equivalent to a 1 per cent. sample solution). METHOD 2.DETERMINATION OF TUNGSTEN AND OF MOLYBDENUM WHEN THE CONCENTRATION OF TUNGSTEN IS HIGHER THAN 0.5 PER CENT.- Preparation of sample solutions-Weigh 1.0000 g of sample into a 250-ml beaker, add 50 ml of the acid mixture A and heat gently. When the sample has dissolved, oxidise the solution by adding nitric acid (sp.gr. 1.42) dropwise, and evaporate the solution until the first fumes of perchloric acid appear. Cool, dilute to 50 ml, filter the solution through a Whatman No. 541 filter-paper and dilute to 100ml.832 THOMERSON AND PRICE: A COMPREHENSIVE SCHEME FOR THE ANALYSIS OF [Analyst, Vol. 96 Preparation of calibration solutions-To each of seven 250-ml beakers, add 1.0 g of pure iron and suitable volumes of stock solutions. Use the dissolution procedure specified above for the samples.EXTENSION OF CALIBRATION RANGES- When it is required to determine higher concentrations of alloying elements than those provided for by the recommended calibration ranges, the sample solution should be diluted by an appropriate factor such that the final solution still contains 1 per cent. of iron. All observations of freedom from inter-elemental interferences were made on solutions containing 1 per cent. of iron. I t may therefore be taken as a general principle that, provided both samples and standards contain 1 per cent. of iron, it is possible to use any degree of dilution. In preparing samples known to require range extension the appropriate amount of 5 per cent. stock iron solution can be added before finally making up to volume. A sample already made up to volume can simply be diluted as required with 1 per cent.iron solution derived by appropriate dilution of the 5 per cent. stock solution. ANALYSIS- Recommended instrumental conditions for common constituent elements are summarised in Table 11. The blank (“zero” standard) and calibration solutions should be aspirated followed by the sample solutions. For highest accuracy the blank solution should be run between all other standards and samples and a high standard repeated after every five or ten samples. The calibration graph of absorbance versus concentration is plotted for each element and the concentrations of the elements in the sample solutions are read off. TABLE I1 INSTRUMENTAL CONDITIONS Mn Ni Cr Mo Cu V Co Ti Sn A1 Pb W Wavelength/nm 279.5 232.0 357.9 313-3 324.8 318.4 240.7 364.3 224.0 309.3 217.0 255.1 Slit width/mm 0.03 0.05 0.05 0.05 0.05 0.05 0.03 0.05 0.10 0-05 0.05 0.05 Burner A A N N A N A N N N A N Observation heightlcm 0.8 0-8 0.5 0.5 0.8 0.7 0.8 0.7 0.7 0.7 0.8 0.8 5.0 - - - Air/l min-l 5.0 5.0 - - 6.0 - 5.0 - Acetylene/l min-l* 1.4 0.8 4.2 4.7t 1.0 4.5 1.2 4.5 4.4 4.2 1.2 4-5 Nitrous oxide/l min-l - - 5.0 5.0 - 5.0 - 5.0 5.0 5.0 - 5.0 A = 10 cm air - acetylene burner.N = 5 cm nitrous oxide - acetylene burner with 0.59-mm jaw width (see Apparatus). * These values should be used as a guide only and the flow-rate must be adjusted to give the best sensitivity for each element. t This acetylene flow-rate is critical and must be carefully adjusted to give a flame with a maximum red “feather” height without luminescence.For the chromium range 25 to 300 mg I-’ the burner must be in the fully rotated position to reduce the sensitivity (for the SP90 atomic-absorption spectrophotometer this is about 50” from the optical axis). DISCUSSION The use of perchloric acid as the final medium was adhered to because it has the advantage that it causes the least interference of all common acids used. Originally proposed for the determination of chromium and molybdenum,l its use has proved ideal for the determination of the remaining elements attempted here. The only addition necessary to the aqueous calibration solutions throughout the scheme is that of iron. Therefore, calibration solutions of mixed elements can be used whenever possible with only one simple addition of the stock iron solution or prepared from I-g portions of pure iron (Proposed method).Titanium, however, can be satisfactorily determined by direct comparison with aqueous standards, but can equally acceptably be incorporated into mixed element standards containing iron. Most of the results given in Tables I11 and IV were obtained by using only one calibration range for each element. If, therefore, a better choice of calibration range were made for specific samples greater precision should be possible in most instances.December, 19711 A WIDE RANGE OF STEELS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 833 TABLE I11 RESULTS OF STEEL ANALYSIS BY USING METHOD 1 Element Manganese Nickel Chromium Molybdenum Copper Vanadium Cobalt Titanium Tin Aluminium (acid soluble) Lead Steel type Mild Permanent magnet alloy Ferritic stainless Stainless Mild Ferritic stainless Stainless Permanent magnet alloy Mild Low alloy Ferritic stainless Austenitic stainless Stainless Mild Mild Low alloy Austenitic stainless Mild Ferritic stainless Permanent magnet alloy Permanent magnet alloy Stainless Stainless Mild High speed Mild Stainless Mild High speed Mild Stainless Permanent magnet alloy Permanent magnet alloy High speed Mild Mild Mild Permanent magnet alloy Permanent magnet alloy Mild Mild B.C.S.No. 321 312 341 23512 32 1 341 23512 233 321 25711 339 334 341 32 1 324 2521 1 336 329 341 312 233 261 23512 329 326 23512 329 321 23512 233 312 24111 324 329 321 233 312 326 329 22011 22011 Value found, per cent.0.126, 0-126, 0.124 0.208, 0.208, 0.208 0-425, 0.425, 0.425 0.90, 0.90, 0.88 0.102, 0.102, 0.098 0.58, 0.59, 0.59 9.6, 9.4, 9-4 11.4, 11-4, 11.4 0.12, 0.11, 0.11 2-98, 2-88, 2.94 12.4, 12.4, 12.5 25.4, 25.7, 25.4 24.3, 23.9, 23.7 0.07, 0.07, 0-07 0-17, 0-17, 0.17 1.08, 1.10, 1.08 2-36, 2.40, 2.36 0.070, 0.070, 0.068 0-12, 0.12, 0-12 3.12, 3-12, 3-12 5.13, 5.11, 5-08 0.02, 0.03, 0.03 0-04, 0.05, 0.04 0.10, 0.08, 0.09 2-05, 1.98, 1.98 -0.02, 0.02, 0.02 0.06, 0.06, 0.06 0.05, 0.05, 0.05 0.12, 0.12, 0.12 0.13, 0-12, 0.12 0.31, 0.30, 0-31 0.83, 0.82, 0.83 1-21, 1-19, 1-17 0.02, 0.03, 0.03 0-13, 0.13, 0.12 0.06, 0.06, 0.06 0.13, 0.13, 0.14 6.96, 6.87, 6.87 7.78, 7.84, 7-92 0.016, 0.016, 0.016 0.050, 0.050, 0-048 TABLE IV Certificate value, per cent. 0.13 0.20 0.43 0.89 0.099 0.56 9.38 11.22 0.11 2.97 12.4 25.6 24.0 0.068 0.17 1.11 2.43 0.072 0.10 3.10 5.09 0.03 0.04 0.083 2-09 0.023 0.056 0.07 0.13 0-13 0.32 0.79 1.19 0.025 0.13 0.053 0.12 6.98 7.87 0.014 0.050 RESULTS OF STEEL ANALYSES BY USING METHOD 2 Certificate Element Steel type B.C.S.No. per cent. per cent. Value found, value, Tungsten Mild High speed High speed Molybdenum High speed High speed 323 0-25, 0.25, 0.25 0.25 22011 6.88, 6.88, 6.88 6-86 24111 19.4, 19.8, 19.4 19.61 24111 0.56, 0.56, 0-56* 0-52 22011 5.30, 5.30, 5-30? 5.20 Certificate range, per cent. 0.12-0.13 0.19-0.22 0.41-0.44 0.88-0.90 0.096-0.105 9-34-9.42 11.14-11*28 0.1 0-0.1 1 2.95-3.01 12.3-12-5 25.5-25- 7 23.9-24.1 0*064-0-072 0.55-0.58 0.16-0.18 1 -09- 1.1 3 2.39-2.46 0.07 1-0.075 2.99-3-15 - 5.04-5.16 - 0.03-0.04 0*079-0.085 2.08-2.13 0.02 1-0.025 0.056-0.057 0.068-0.072 0.12-0.13 0.127-0.139 0.3 1-0.33 0.78-0.80 1 - 16-1-23 0.024-0*028 0.12-0- 14 0.049-0.057 6.89-7-08 7-7 7-7-93 0.01 2-0.0 17 0.042-0.052 - Certificate range, per cent.0.24-0.26 19.561 9.72 6.7 8-7-00 0.51-0.52 5.15-5.2 7 * Results obtained in presence of 19-6 per cent. of tungsten. t Results obtained in presence of 6.9 per cent. of tungsten.834 THOMERSON AND PRICE Based on the fact that interference and suppressions are minimised, and between samples and standards are equalised, in the presence of 1 per cent. of iron, the method is readily ex- tended to higher concentration ranges simply by the dilution of samples and addition of iron to give a concentration of 1 per cent. An increase in reading accuracy can also be achieved by using the difference method that has previously been described.1937 The only restriction on the use of perchloric acid results from its failure to retain silicon and tungsten in solution.The tungsten can be determined separately (Method 2) without any problem arising. The silicon recovery, after filtering the sample solution, is quantitative, so we recommend the use of either the traditional simple gravimetric method or a separate atomic-absorption spectrophotometric method of analysis as exemplified by Price and Roos38 and by M~Auliffe.~Y A great advantage most atomic-absorption methods have over traditional techniques is the small sample volume or weight required. This scheme for the analysis of steels is no exception and it is possible to carry out the complete analysis on only 1 g of sample (for the determination of tungsten and silicon separate sample weighings are, of course, needed).Consequently much time can be saved compared with most laboratory methods in common use. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. REFERENCES Thomerson, D. R., and Price, W. J., Analyst, 1971, 96, 321. Belcher, C. B., and Kinson, K., Analytica Chim. Acta, 1964, 30, 483. Atsuya, I., Sci. Rep. Res. Insts Tdhoku Univ., A , 1967, 19, 59. Ramirez-Munoz, J., and Roth, M. E., Beckman Flame Notes, 1969, 2, 3. Hubbard, D. P., and Monks, H. H., Analytica Chim. Acta, 1969, 47, 197. McPherson, G. L., “Proceedings of the 17th BISRA Chemists Conference, Scarborough, 1964,” Suzuki, M., and Takeuchi, T., J .Chem. SOC. Japan, Ind. Chem. Sect., 1964, 67, 1207. Beyer, M., Atomic Absorption Newsletter, 1965, 4, 212. Sprague, S., and Slavin, W., Dev. Appl. Spectrosc., 1965, 4, 433. Clarke, W. E., British Cast Iron Research Association Report No. 873, 1967, p. 243. Price, W. J., “XIIIth Colloquium Spectroscopicum Internationale,” Ottawa, 1967. Clarke, W. E., and Cooke, P. A., British Cast Iron Research Association Report No. 891, 1967, Price, W. J., and Cooke, P. A., Spectrovision, 1967, 18, 2. Heinz, K., and Ohls, K., Arch. EisenhtiittWes., 1968, 39, 925. Jimenez Seco, J. L., and Gomez Coedo, A., Revta Metalurgia, 1968, 4, 621. Knight, D. M., and Pyzyna, M. K., Atomic Absorption Newsletter, 1968, 8, 129. Feldman, F. J., Blasi, J. A., and Smith, S. B., Analyt. Chem., 1969, 41, 1095. Kinson, K., and Belcher, C. B., Analytica Chim. Acta, 1964, 30, 64. Ramirez-Munoz, J., and Roth, M. E., Beckman Flame Notes, 1969, 4, 102. Carper, J. L., Atomic Absorption Newsletter, 1970, 9, 2. Wallace, F. J., Foseco Dev., 1961, 7, 54. -, Hilger J., 1963, 7, 65. Kinson, K., and Belcher, C. B., Analytica Chim,.. Acta, 1964, 31, 180. Konig, P., Heinz, K., and Thieman, E., Arch. EisenhiittWes., 1969, 40, 53. Capacho-Delgado, L., and Manning, D. C., Atomic Absorption Newsletter, 1966, 5, 1. McPherson, G. L., Price, J. W., and Scaife, P. H., Nature, Lond., 1963, 199, 371. Lockyer, R. L., Hilger Watts Research Report, BR.26, 1965. Bowman, J. A., and Willis, J. B., Analyt. Chem., 1967, 39, 1210. Headridge, J. B., and Hubbard, D. P., Analytica Chim. Acta, 1967, 37, 161. Mostyn, R. A., and Cunningham, A. F., Atomic Absorption Newsletter, 1967, 6, 86. Nikolaev, G. I., J . Analyt. Chem. U.S.S.R., 1965, 20, 412. Amos, M. D., and Thomas, P. E., Analytica Chim. Acta, 1965, 32, 139. Endo, Y., Ohata, H., and Nakahara, Y., Japan Analyst, 1967, 16, 364. Konig, P., Schmitz, K. H., and Thieman, E., 2. analyt. Chem.. 1969, 244, 232. Elwell, W. T., and Gidley, J. A. F., Analytica Chim. Acta, 1961, 24, 71. Dagnall, R. AT., West, T. S., and Young, P., Analyt. Ckem., 1966, 38, 358. Thomerson, D. R., Spectrovision, 1971, 25, 12. Price, W. J., and Roos, J. T. H., Analyst, 2968, 93, 709. McAuliffe, J. J., Atomic Absorption Newsletter, 1967, 6, 69. 1964, p. 12. p. 526. Received February 3rd, 197 1 Accepted July 30th, 197 1