- Email: [email protected]
Determination of aluminium in iron, steel and ferrous and non-ferrous alloys by atomic-absorption spectrophotometry after a mercury-cathode separation and extraction of the aluminium—acetylacetone complex
0039-9l40/8l/O70461-07102.00/0 Pergamon Press Ltd
Talanru. Vol. 28. pp. 461 to 467. 1981 Printed m Great Br~tam
DETERMINATION OF ALUMINIUM IN IRON, STEEL AND FERROUS AND NON-FERROUS ALLOYS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY AFTER A MERCURY-CATHODE SEPARATION AND EXTRACTION OF THE ALUMINIUMACETYLACETONE COMPLEX ELSIE M. DCINALD~~N Mineral Sciences Laboratories, Canada Centre for Mineral and Energy Technology, Department of Energy, Mines and Resources, Ottawa, Canada (Received 10 December 1980. Accepted I January 1981)
Summary-A method for determining 0.0005% or more of total aluminium in high- and low-alloy steels, iron and ferrovanadium is described. Iron, chromium and other matrix elements are separated from aluminium by electrolysis with a mercury cathode and aluminium is separated from tungsten, titanium, vanadium and phosphate by chloroform extraction of its acetylacetone complex at pH 6.5 from an ammonium acetate-hydrogen peroxide medium. The extract is evaporated to dryness and organic material is destroyed with nitric and perchloric acids. Aluminium is determined by atomic-absorption spectrophotometry in a nitrous oxide-acetylene flame, at 309.3 nm, in a 5% v/v perchloric acid medium containing 1000 pg of sodium per ml. Acid-soluble and acid-insoluble aluminium can also be determined. The method is also applicable to copper- and nickel-base alloys. Results obtained by this method are compared with those obtained spectrophotometrically with Pyrocatechol Violet, after the separations described above followed by the separation of the residual co-extracted iron and copper by a combined ammonium pyrrolidinedithiocarbamate-cupferron-chloroform extraction from 10% v/v hydrochloric acid medium.
A reasonably simple and reliable flame atomicabsorption spectrophotometric (AAS) method for the determination of -0.0005-0.05% of total aluminium in both low- and high-alloy steels was required for use in routine work in the CANMET chemical laboratory. Because of the low sensitivity of AAS for aluminium and because iron interferes in the determination in a nitrous oxide-acetylene flame,‘-’ most previous investigatorsG6 agree that a separation from iron and a concentration step are required for the accurate determination of
ium in both high-alloy and mild steels is based on methyl isobutyl ketone extraction of the iron(II1) chloro-complex and the subsequent separation of aluminium from the bulk of the alloying constituents mentioned above, by extracting it as the acetylacetone complex from a weakly acidic medium with pure acetylacetone.‘j After concentration of the extract by . evaporation, aluminium is determined by spraying the solution into the flame. In this method, an inordinately large volume of acetylacetone is used for the extraction of aluminium, and the residual iron and many other elements are co-extracted under the conditions employed. About ten years ago, the author developed a spectrophotometric method, based on complex formation with Pyrocatechol Violet, for the determination of small amounts of aluminium in iron, steel and other materials.’ This method also involves the separation of aluminium by extraction of its acetylacetone complex, after the preliminary separation of iron, chromium, nickel, molybdenum and various other elements by electrolysis with a mercury cathode. However, the selectivity of the acetylacetone extraction step was increased by extracting the aluminium complex into chloroform from 0.05M acetylacetone medium in the presence of hydrogen peroxide as a complexing agent for molybdenum, tungsten, titanium and vanadium. The residual co-extracted iron and small amounts of other co-extracted elements
ELSIEM. DONALDSON
462
were subsequently separated from aluminium by a combined ammonium pyrrolidinedithiocarbamatecupferron-chloroform extraction from a 10% hydrochloric acid medium, after stripping of the complex from the chloroform phase into concentrated hydrochloric acid. Because the mercury-cathode separation in conjunction with the acetylacetone extractionpre-concentration step described in this method produces a solution that is essentially free of matrix elements except for a small amount of iron, it was considered that an AAS finish could readily be applied after the extraction step. This paper describes the successful determination of total aluminium or both acid-soluble and acid-insoluble aluminium in mild and high-alloy steels and in ferrovanadium. Total aluminium can also be determined in copper- and nickel-b&e alloys. After a preliminary mercury-cathode separation of matrix elements from a 2-3% v/v perchloric or sulphuric acid medium, aluminium is separated from the remaining elements, except residual iron or copper, by chloroform extraction of its acetylacetone complex at pH 6.5 from a O.lM acetylacetone-ammonium acetatehydrogen peroxide medium. The extract is evaporated to dryness and the residue is treated with nitric and perchloric acids. Results obtained by this method are compared with those obtained previously by the spectrophotometric Pyrocatechol Violet method.’ EXPERIMENTAL
A Varian-Techtron model AA6 spectrophotometer, equipped with a 6-cm laminar-flow nitrous oxide-acetylene burner and an aluminium hollow-cathode lamp, was used under the following conditions (Note 1). Wavelength : 309.3 nm Lamp current: 10 mA Spectral band-pass: 0.20 nm Height of light-path above burner: Acetylene flowmeter reading: 6.75 Nitrous oxide flowmeter reading: 5 1.5-2 cm Flame: non-luminous, Aspiration rate: 2 ml/min
9 mm (_ 4.5 l./min) 6 (_ 8.5 I./min) “red feather”
Reagents Standard aluminium solution. 2000 &ml. Dissolve 1.000 g of high-purity aluminium metal and about 5 mg of highpurity iron metal (Note 2) by heating gently in a covered &O-ml beaker with 20 ml bf concentrated hydrochloric acid. Cool and dilute to 1 litre with water. Prepare a lOO-pg/ml solution by diluting 20 ml of this stock solution to 200 ml with water. Prepare the diluted solution fresh as required. Acetylacetone, 20% solution. Dilute 20 ml of acetylacetone and 30 ml of ethyl alcohol to 100 ml with water. This solution is stable for at least one week. Sodium 10 mg/ml solution. Dissolve 12.7 g of sodium chloride in water and dilute to 500 ml. Perchloric acid, 50% v/v. Sulphuric acid, 1% v/v. Hydrogen peroxide, 30%. Ammonium acetate solution, 50%. Procedures Calibration solutions. Add 10 ml each of 50% perchloric acid and lO-mg/ml sodium solution to each of nine 100-m]
standard flasks; then, from a burette, add to the first eight flasks 1, 2, 3, 5, 10, 15, 20 and 25 ml respectively, of the lOO-pgg/ml standard aluminium solution. The last flask contains the zero calibration solution. Dilute each solution to volume with water (Note 3). Iron, steel and nickel-chromium alloys. Transfer 0.1-l g of sample (Note 4) containing not more than 2 mg of aluminium to a 400-ml beaker, then cover the beaker and add 15 ml of concentrated hydrochloric acid and 5 ml of concentrated nitric acid. Heat until the sample has dissolved, then add 20 ml of concentrated perchloric acid (Note 5) and evaporate the solution to fumes of perchloric acid. Remove the cover and carefully evaporate the solution to 5-7 ml. Cool, add -50 ml of water and heat to dissolve the salts. Add about a quarter of a Whatman filter pulp tablet, macerate it with a stirring rod, then filter the solution (Whatman No. 40 paper) into a 400-m] beaker and transfer the residue quantitatively to the filter paper. Wash the paper and the residue three times with water, then three times with 1% sulphuric acid to ensure the complete removal of perchloric acid. Evaporate the filtrate to about 75 ml (Note 6). Carry a blank through the whole procedure. Transfer the paper containing the residue to a 30-ml platinum crucible, burn off the paper at a low temperature and ignite at 600-700”. Cool the crucible and add 1 ml each of 50% sulphuric acid and concentrated nitric and hydrofluoric acids. Heat gently to dissolve the residue, then evaporate the solution to dryness. Fuse the residue with 2 g of fused sodium hydrogen sulphate (Note 7), then cool the crucible and transfer it to a 250-ml beaker containing 4 50 ml of water. Cover the beaker and heat gently to dissolve the melt. Remove the crucible after washing it thoroughly with water (Note 8), then add the solution to the initial filtrate. Transfer the solution to a mercury-cathode cell, dilute to -200 ml with water and electrolyse for 75 min at -_ 10 A. Filter the electrolyte (Whatman No. 541 paper) into the 400-ml beaker that initially contained the solution. Wash the cathode cell and the paper thoroughly with water, then discard the paper. Add 5 ml of concentrated hydrochloric acid and 3 ml of 50”% sulphuric acid and evaporate the solution until 1.5-2 ml of sulphuric acid remain (Note 9). Cool the beaker, add 50 ml of water, heat gently until the solution is clear (Note lo), then cool it to room temperature. In succession, add 2 ml of 30% hydrogen peroxide, 5 ml of 20% acetylacetone solution and 10 ml of 50% ammonium acetate solution to the solution, then, using a pHmeter, adjust the pH of the solution to 6.5 f 0.1 with concentrated ammonia solution. Transfer the solution to a 125-ml separatory funnel, add 10 ml of chloroform, stopper and shake for 2 min. Allow several min for the layers to separate, then drain the chloroform phase into a 100-ml beaker. Extract the aqueous phase twice more by shaking it for 1 min each time with 5 ml of chloroform, then wash it by shaking it for -30 set with 5 ml of chloroform. Add 5 drops of concentrated hydrochloric acid to the combined extracts and evaporate to dryness in a hot water-bath. Add 1 ml of 50% perchloric acid and 2 ml of concentrated nitric acid, then cover the beaker and heat on a hot plate until fumes of perchloric acid are evolved (Note 11). Cool the solution to room temperature and add -3 ml of water. Transfer the blank solution to a lo-ml standard flask containing 1 ml of lo-mg/ml sodium solution and dilute to volume with water. Depending on the expected aluminium content, transfer the sample solution to a standard flask of appropriate size (lo-100 ml) and add sufficient lo-mg/ml sodium solution for 1 ml to be present for each 10 ml of final solution. If necessary, add sufficient 50% perchloric acid for 1 ml to be present for each 10 ml of final solution in excess of 10 ml (Note 12). and dilute the solution to volume with water. Measure the absorbances at 309.3 nm when the resulting
463
Determination of aluminium in iron, steel and alloys solutions are aspirated into a strongly reducing nitrous oxide-acetylene flame (Note 13). Calculate the aluminium contents (in mg) from the sample and blank absorbances and those obtained concurrently for calibration solutions that bracket the sample and blank concentrations. Correct the result obtained for the sample solution by subtracting that obtained for the blank solution. Ferrouanadium.Decompose 0.1-0.5 g of sample containing up to 2 mg of aluminium as described above, using 20 ml of 50% sulphuric acid instead of concentrated perchloric acid. Heat the solution until the evolution of oxides of nitrogen ceases, then remove the cover and evaporate to 5-7 ml. Cool, add 50 ml of water, heat to dissolve the salts, then proceed with the filtration of the solution (Note 14) the treatment of the residue, the mercury-cathode separation (Note l5), the extraction (Note 16) and the subsequent determination of aluminium as described above. Copper-baseand nickel-copper alloys.Transfer 0.1-l g of sample containing up to 2 mg of aluminium to a 400-ml beaker (Note 17). then cover the beaker and add 5 ml each of water and concentrated nitric acid and 20 ml of SOY0 sulphuric acid. Heat the solution until the evolution of oxides of nitrogen ceases, then remove the cover and evaporate to 5-7 ml. Cool, add 50 ml of water and, if necessary, heat to dissolve the salts. Transfer the solution to an electrolytic cell, dilute to 200 ml with water and proceed with the mercury-cathode separation (Note IS), the extraction and the subsequent determination of aluminium, as described above. Notes 1. A strongly reducing, non-luminous nitrous oxideacetylene flame is required to obtain the highest sensitivity for aluminillm. The height at which the beam from the hollow-cathode lamp passes through the flame is also extremely important. 1.3 Consequently, after all other instrumental parameters have been set, the acetylene and nitrous oxide flow-rates should be adjusted to give the maximum absorbance when a solution containing aluminium is aspirated into the flame. 2. Iron hastens the dissolution of the aluminium metal. 3. The calibration solutions should be prepared fresh every week. 4. More than 1 g of sample is not recommended because too much iron may remain in the solution after the mercury-cathode separation. The residual iron is co-extracted and, if a large amount is present, may result in an explosive reaction when the final solution is evaporated to fumes of perchloric acid. Too much iron can also cause a slight positive error in the aluminium result. 5. For iron and steel of low chromium content and high titanium, vanadium, molybdenum or tungsten content, add 20 ml of 50% sulphuric acid instead of perchloric acid and proceed as described for ferrovanadium. 6. If the acid-soluble aluminium content of the sample is required, proceed with the mercury-cathode separation, the extraction and the subsequent determination of aluminium as described. 7. One gram of a mixture consisting of 75% by weight of sodium carbonate and 25% by weight of fused boric acid can also be used for fusion of the insoluble material.4 However, the melt should be dissolved in water containing _ 2 ml of 50% sulphuric acid. 8. If the acid-insoluble aluminium content of the sample is required, it can be determined in the resulting solution. In this case, omit the mercury-cathode separation, add the volumes of hydrogen peroxide, acetylacetone and ammonium acetate solutions recommended in the subsequent part of the procedure, and proceed with the pH-adjustment, the extraction and the determination of aluminium as described. The total amount of aluminium present in the sample can then be determined by adding that found in the
acid-insoluble material to that found in the filtrate (Note 6). The separate determination of acid-insoluble aluminium is recommended if an appreciable amount of tungsten trioxide is present in the acid-insoluble material. In this case, add 5 ml of 30% hydrogen peroxide and the recommended volumes of acetylacetone and ammonium acetate solutions to the solution obtained after dissolving the melt, then add sufficient concentrated ammonia solution to dissolve the tungsten trioxide. Acidify the solution with 50% sulphuric acid, adjust the pH to 6.5 f 0.1 and proceed with the extraction and subsequent determination of aluminium. 9. This step volatilizes any mercury left in the solution. Sulphuric acid solutions of aluminium should not be evaporated to dryness. This results in the formation of an anhydrous aluminium sulphate that is virtually insoluble in water or in dilute acid and causes a low result for aluminiurn.’ 10. If the extraction cannot be completed the same day, allow the solution to stand overnight at this point. 11. The solution should not be allowed to evaporate to dryness. If the atomic-absorption determination cannot be completed the same day, allow the solution to stand overnight at this point. The final aluminium solution is not always stable on prolonged standing. 12. Additional 50% perchloric acid is not required if the final volume of the solution is to be 10 ml, because 1 ml is added to the residue remaining after the removal of chloroform. For final sample solution volumes of 25, 50 or 100 ml, add 1.5, 4 or 9 ml, respectively. 13. Between 5- and lO-fold scale expansion is recommended for the determination of aluminium. 14. It is not necessary to wash the paper and residue with 1% sulphuric acid as described in the procedure for iron and steel. 15. It is not necessary to add 3 ml of 50% sulphuric acid after the mercury-cathode separation. 16. More than 2 ml of 30% hydrogen peroxide-as recommended in the procedure for iron and steel-will be required for samples of high vanadium content. Approximately I ml is required to complex 30 mg of vanadium. Five millilitres should be added for a 0.5-g sample of ferrovanadium containing about 30% of vanadium. The solution should be yellow-not green-after the pH-adj stY ment. 17. If the sample contains silicon, use a 400-ml Teflon beaker and add 1 or 2 ml of concentrated hydrofluoric acid during the decomposition step. RESULTS
Atomic-absorption finish after separation of aluminium by extraction of its acetylacetone complex
Tests showed that, after the separation of aluminium by extraction of its acetylacetone complex into chloroform at pH 6.5 as described earlier,’ aluminktm can readily be determined by AAS in a nitrous oxideacetylene flame after evaporation of the extract and a small amount of concentrated hydrochloric acid to dryness and destruction of the organic material in the residue with nitric and perchloric acids as described in the proposed method. At least 0.5 ml of concentrated perchloric acid is required for this decomposition step. Consequently, a perchloric acid concentration of 5% by volume was chosen for the sample and calibration solutions because a final sample solution volume of at least 10 ml is necessary for the determination of small amounts of aluminium. To
464
ELSIE M. DONALDSON
suppress the ionization of aluminium, 1000 pg of sodium (as the chloride) per ml was also added to the sample and calibration solutions.’ EfSpct of diverse ions The electrolyte obtained after the separation of iron, chromium and other matrix elements from a dilute perchloric or sulphuric acid solution of an iron or steel sample by electrolysis with a mercury-cathode would contain the titanium, zirconium, vanadium and phosphorus (as phosphate) present in the sample, some of the manganese and tungsten and small amounts of residual iron, chromium, molybdenum and possibly nickel, cobalt and copper. It was shown previously’ that tungsten, niobium and tantalum can be kept in solution, and that molybdenum, titanium and vanadium can be prevented from co-extracting as acetylacetone complexes, by complexing them with hydrogen peroxide during the extraction of the aluminium-acetylacetone complex into chloroform at pH 6.5 from a 0.05M acetylacetone medium. Under these conditions, only iron(III), copper(I1) and beryllium are completely co-extracted. Cobalt(II), lead and chromium(III), which is produced in an acid medium containing hydrogen peroxide, are very slightly coextracted, and more than about 8 mg of phosphate causes a low result for aluminium because of the formation of aluminium phosphate. Tests, in which aluminium was extracted from a O.lM acetylacetone medium to increase the range of the proposed method, showed that up to at least 10 mg of iron(II1) or 5 mg of copper(I1) will not interfere in the extraction of up to at least 2 mg of aluminium. Furthermore, these quantities of iron and copper will not cause significant error in the aluminium result when the final solution is diluted to 10 ml. A larger amount of iron can cause a slight positive error ( - 3% at the lOO+g level) in the result. However, analysis of the final sample solutions derived from l-g samples of iron and steel showed that usually less than 5 mg of iron will remain in the electrolyte after the mercurycathode separation. Up to at least 10 mg of cobalt(II), lead, nickel and chromium(III), 20 mg of manganese(I1) and phosphate, and 5 mg of niobium and tantalum can be present during the extraction step without interfering in the extraction or subsequent determination of aluminium. In the presence of 2 ml of 30% w/v hydrogen peroxide, up to at least 60 mg of vanadium(V) and 20 mg of molybdenum(VI), tungsten(V1) and titanium(IV) will not interfere in the extraction of aluminium. Interference from up to at least 150 mg of vanadium can be avoided by adding 5 ml of 30% hydrogen peroxide before the pH-adjustment. Any insoluble tungsten trioxide present at this stage can be dissolved by adding concentrated ammonia solution. Subsequently, the solution should be acidified with dilute sulphuric acid followed by the adjustment of the pH to 6.5 + 0.1 with concentrated ammonia solution.
The effect of beryllium on the determination of aluminium by AAS was not investigated because it is not usually present in iron and steel. Although it was found previously’ that small amounts of zirconium can interfere in the extraction of aluminium by causing emulsification in the chloroform phase, up to at least 0.5 mg will not interfere in the proposed method. Applications To test the reliability of the proposed method, it was applied to the analysis of National Bureau of Standards (NBS) and British Chemical Standards (BCS) certified reference iron, steel and ferrovanadium samples and to a stainless steel to which known amounts of aluminium were added. It was also applied to NBS copper- and nickel-base alloys. The results of these analyses, which are the means of four or five AAS measurements, are given in Table 1. All the results shown are for individual samples. The results obtained previously for some of these reference materials by the spectrophotometric Pyrocatechol Violet method’ are also given in Table 1. Precision The precision at about the 0.0005 and the 0.002% levels of aluminium was tested by analysing one series of five samples of a BCS high-silicon steel for both acid-soluble and acid-insoluble aluminium. A second series of five samples was analysed for total aluminium. Although this reference material has not been certified for aluminium, results for acid-soluble, acidinsoluble and total aluminium have been reported by previous investigators. 4.5*‘o The results of these tests, which are the means of four or five AAS measurements on each sample solution, are given in Table 2.
DISCUSSION
Table 1 shows that the results obtained for the NBS and BCS iron, steel, ferrovanadium and non-ferrous alloys are in good agreement with the certified values and, where applicable, with the results obtained previously by the Pyrocatechol Violet method.’ The results obtained for the stainless-steel sample to which known amounts of aluminium were added, agree with the total calculated amount present. Table 2 shows that the relative standard deviation for acid-soluble aluminium at the 0.0004~~ level-and consequently also for total aluminium-is relatively high. This is partly due to the magnitude of the reagent blank, which is given below. At the 0.002% level, the precision of the results for acid-insoluble aluminium, for total aluminium determined by adding acid-insoluble and acid-soluble aluminium and for total aluminium determined in one step is comparable with that obtained by Shaw and Ottaway” by flameless AAS and with that obtained by Jenkins and Jones5 and other investigators using conventional
Determination of aluminium in iron, steel and alloys
465
E a~
o
e,¢
o., eq-
a~
E O
rz O
Z
¢.
g G
.E
o
~g
e-
llt
=
I I
1
2
g~ o
0
Z .E
Z
E
.o
E
E
._~ .E N
o
~,,,E
a
. .O.
.~_
Z
l
~
UN
~N
~NUo
E
•
~
~
N ~ ..~ ++
I--
.E~ E
±i
~.~.~
E~
~ m ZZZZZ
~'~ ZZ
Z
ZZmm~mZZZZZ
ZZ
466
ELSIEM. DONALDSON
AAS.4*6 The precision of the results obtained for total aluminium determined in one step is better than that obtained when acid-soluble and acid-insoluble aluminium are determined separately and added. However, this is expected from statistical considerations and because of the fewer manipulations required. Table 2 also shows that the results obtained for acid-soluble, acid-insoluble and total aluminium by the proposed method are in excellent agreement with those obtained by previous investigators. Although sodium hydrogen sulphate is recommended for fusion of the acid-insoluble material in methods for determining total aluminium in iron, steel and various other materials,ll~lZ Cobb et aL4 found that the use of this flux produced variable and low results for aluminium in several BCS mild steel samples. Consequently, they recommended a mixture of sodium carbonate and boric acid. The particular reference materials used in their tests were not immediately available. However, the results obtained for acid-insoluble and total aluminium in the BCS highsilicon steel by the proposed method (Table 2) show that sodium hydrogen sulphate is a suitable reagent for fusion purposes. This is also indicated in Table 1. The results obtained for NBS samples 19g, 55e and 1OOA by the proposed method involving the use of sodium hydrogen sulphate for fusion of the acidinsoluble material are in excellent agreement with those obtained (shown in brackets) when a mixture of sodium carbonate and fused boric acid was used for fusion purposes. The proposed method has some definite advantages over the recent AAS method6 mentioned earlier, which is also applicable to high-alloy steels, and which involves a preliminary separation of iron by methyl isobutyl ketone extraction of the chloro-complex. In this method, the sample, including the acidinsoluble material, is decomposed in a Teflon-lined (85-ml capacity cup) pressure bomb. Aluminium is ultimately separated from the bulk of the alloying constituents such as chromium and nickel by a double extraction of its acetylacetone complex into 50 ml of pure acetylacetone which, because of its low density, requires the use of two separatory funnels. The combined acetylacetone extracts are then concentrated to -7 ml by evaporation in an oil-bath at 150-160” and aluminium is determined by spraying the extract into the flame. This requires the use of calibration solutions prepared in a similar manner by extracting known amounts of aluminium into pure acetylacetone. Part of the manganese, lead, cobalt and nickel present in the sample, much of the niobium, titanium and vanadium, most of the copper, tin and zirconium, and all of the iron(M) remaining after the methyl isobutyl ketone extraction step will be coextracted with the aluminium. In the proposed method for iron and steel, 1 ml of acetylacetone is used and the aluminium complex is extracted into chloroform which, because of its high density, is convenient to use when multiple extractions are required.
Determination
of aluminium in iron, steel and alloys
The extract is easily evaporated to dryness in the presence of a small amount of concentrated hydrochloric acid, and aluminium is ultimately determined in an aqueous medium which facilitates the preparation of calibration solutions. After a mercury-cathode separation, and in the presence of hydrogen peroxide as a complexing agent, essentially only the residual iron and a small amount of chromium(II1) are co-extracted with the aluminium. As stated previously, at the l-g level, usually less than 5 mg of iron(II1) will remain in the electrolyte. Although the pressure-bomb decomposition step described in the Headridge and Sowerbutts method6 would be advantageous for the determination of total aluminium, this was not pursued in this work because of the low capacity (25 ml) of the Teflon cup in the available commercial acid-digestion bombs. Attempts to evaporate the solution obtained after dissolving 1 g of steel with aqua regia to low volume in a 25-ml Teflon cup resulted in considerable loss of sample by spray. The proposed method also has some advantages over the previous AAS methods involving the preliminary separation of iron by isobutyl acetate extraction of the chloro-complex.4*5 In these methods, _ 50-150 mg of iron remain in the aqueous phase after the extraction. This requires the addition of iron to the calibration solutions to eliminate error. In the proposed method, so little iron remains in the electrolyte after the mercury cathode separation that the addition of iron to the calibration solutions is not necessary. The results shown in Table 1 for the copper-base alloys were obtained after ultimately separating aluminium by acetylacetone extraction. However, tests showed that at the l-g copper level, less than 0.5 mg remains in the electrolyte after the mercury-cathode separation. Consequently, the extraction step would not usually be necessary for copper-base alloys, and probably not for most nickel-base alloys, unless appreciable titanium, tungsten or vanadium is present. However, the mercury-cathode separation should be carried out in a perchloric acid medium
TAL. 28/7A--0
467
instead of a sulphuric acid medium and the electrolyte should be evaporated to -0.5 ml if the final solution for the AAS determination is to be 10 ml. The proposed method is suitable for iron and steel containing as little as 0.0005% of aluminium and for copper- and nickel-base alloys containing w 0.0002~0 or more. However, the accuracy that can be obtained at these levels depends on the magnitude of the respective reagent blanks. In this work, the blank varied from c 12 to 20 pg of aluminium during the determination of total aluminium in iron and steel. Approximately two-thirds of the aluminium present in this blank resulted from the sodium hydrogen sulphate used for fusion of the acid-insoluble material. The blanks obtained after fusion with sodium carbonate and fused boric acid were of the same magnitude. Those obtained during the analysis of the non-ferrous materials varied from 4 5 to 8 pg of aluminium.
REFERENCES
1. M. D. Amos and P. E. Thomas, Anal. Chim. Acta, 1965, 32, 139. 2. P. Konig, K.-H. Schmitz and E. Thiemann, Z. Anal. Chem., 1969, 244,232.
3. J. Y. Marks and G. G. Welcher, Anal. Chem., 1970, 42, 1033. 4. W. D. Cobb, W. W. Foster and T. S. Harrison, Lab. Pratt., 1975, 24, 143.
5. R. H. Jenkins and C. P. Jones, Proc. Sot. Anal. Chem., 1972, 9, 266. 6. J. B. Headridge and A. Sowerbutts, Analyst, 1973, 98, 51.
I. E. M. Donaldson, Talanta, 1971, 18, 905. 8. 1. M. Kolthoff and P. J. Elving, Treatise on Analytical Chemistry, Part II, Vol. 4, p. 377. Interscience, New York, 1966. 9. W. Slavin, Atomic Absorption Spectroscopy, p. 79. Interscience, New York, 1968. 10. F. Shaw and J. M. Ottaway, Analyst, 1975, 100, 217. 11. Annual Book of ASTM Standards, Chemical Analysis of Metals; Sampling and Analysis
of Metal Bearing Ores,
Part 12, pp.-549 653, 655,. 692:American Society for Testina and Materials. Philadelnhia. 1979. 12. Z. St&k, P. Povondra and J. Doleial, CRC Crit. Reu. Anal. Chem., 1977, 6, 255.
Talanru. Vol. 28. pp. 461 to 467. 1981 Printed m Great Br~tam
DETERMINATION OF ALUMINIUM IN IRON, STEEL AND FERROUS AND NON-FERROUS ALLOYS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY AFTER A MERCURY-CATHODE SEPARATION AND EXTRACTION OF THE ALUMINIUMACETYLACETONE COMPLEX ELSIE M. DCINALD~~N Mineral Sciences Laboratories, Canada Centre for Mineral and Energy Technology, Department of Energy, Mines and Resources, Ottawa, Canada (Received 10 December 1980. Accepted I January 1981)
Summary-A method for determining 0.0005% or more of total aluminium in high- and low-alloy steels, iron and ferrovanadium is described. Iron, chromium and other matrix elements are separated from aluminium by electrolysis with a mercury cathode and aluminium is separated from tungsten, titanium, vanadium and phosphate by chloroform extraction of its acetylacetone complex at pH 6.5 from an ammonium acetate-hydrogen peroxide medium. The extract is evaporated to dryness and organic material is destroyed with nitric and perchloric acids. Aluminium is determined by atomic-absorption spectrophotometry in a nitrous oxide-acetylene flame, at 309.3 nm, in a 5% v/v perchloric acid medium containing 1000 pg of sodium per ml. Acid-soluble and acid-insoluble aluminium can also be determined. The method is also applicable to copper- and nickel-base alloys. Results obtained by this method are compared with those obtained spectrophotometrically with Pyrocatechol Violet, after the separations described above followed by the separation of the residual co-extracted iron and copper by a combined ammonium pyrrolidinedithiocarbamate-cupferron-chloroform extraction from 10% v/v hydrochloric acid medium.
A reasonably simple and reliable flame atomicabsorption spectrophotometric (AAS) method for the determination of -0.0005-0.05% of total aluminium in both low- and high-alloy steels was required for use in routine work in the CANMET chemical laboratory. Because of the low sensitivity of AAS for aluminium and because iron interferes in the determination in a nitrous oxide-acetylene flame,‘-’ most previous investigatorsG6 agree that a separation from iron and a concentration step are required for the accurate determination of
ium in both high-alloy and mild steels is based on methyl isobutyl ketone extraction of the iron(II1) chloro-complex and the subsequent separation of aluminium from the bulk of the alloying constituents mentioned above, by extracting it as the acetylacetone complex from a weakly acidic medium with pure acetylacetone.‘j After concentration of the extract by . evaporation, aluminium is determined by spraying the solution into the flame. In this method, an inordinately large volume of acetylacetone is used for the extraction of aluminium, and the residual iron and many other elements are co-extracted under the conditions employed. About ten years ago, the author developed a spectrophotometric method, based on complex formation with Pyrocatechol Violet, for the determination of small amounts of aluminium in iron, steel and other materials.’ This method also involves the separation of aluminium by extraction of its acetylacetone complex, after the preliminary separation of iron, chromium, nickel, molybdenum and various other elements by electrolysis with a mercury cathode. However, the selectivity of the acetylacetone extraction step was increased by extracting the aluminium complex into chloroform from 0.05M acetylacetone medium in the presence of hydrogen peroxide as a complexing agent for molybdenum, tungsten, titanium and vanadium. The residual co-extracted iron and small amounts of other co-extracted elements
ELSIEM. DONALDSON
462
were subsequently separated from aluminium by a combined ammonium pyrrolidinedithiocarbamatecupferron-chloroform extraction from a 10% hydrochloric acid medium, after stripping of the complex from the chloroform phase into concentrated hydrochloric acid. Because the mercury-cathode separation in conjunction with the acetylacetone extractionpre-concentration step described in this method produces a solution that is essentially free of matrix elements except for a small amount of iron, it was considered that an AAS finish could readily be applied after the extraction step. This paper describes the successful determination of total aluminium or both acid-soluble and acid-insoluble aluminium in mild and high-alloy steels and in ferrovanadium. Total aluminium can also be determined in copper- and nickel-b&e alloys. After a preliminary mercury-cathode separation of matrix elements from a 2-3% v/v perchloric or sulphuric acid medium, aluminium is separated from the remaining elements, except residual iron or copper, by chloroform extraction of its acetylacetone complex at pH 6.5 from a O.lM acetylacetone-ammonium acetatehydrogen peroxide medium. The extract is evaporated to dryness and the residue is treated with nitric and perchloric acids. Results obtained by this method are compared with those obtained previously by the spectrophotometric Pyrocatechol Violet method.’ EXPERIMENTAL
A Varian-Techtron model AA6 spectrophotometer, equipped with a 6-cm laminar-flow nitrous oxide-acetylene burner and an aluminium hollow-cathode lamp, was used under the following conditions (Note 1). Wavelength : 309.3 nm Lamp current: 10 mA Spectral band-pass: 0.20 nm Height of light-path above burner: Acetylene flowmeter reading: 6.75 Nitrous oxide flowmeter reading: 5 1.5-2 cm Flame: non-luminous, Aspiration rate: 2 ml/min
9 mm (_ 4.5 l./min) 6 (_ 8.5 I./min) “red feather”
Reagents Standard aluminium solution. 2000 &ml. Dissolve 1.000 g of high-purity aluminium metal and about 5 mg of highpurity iron metal (Note 2) by heating gently in a covered &O-ml beaker with 20 ml bf concentrated hydrochloric acid. Cool and dilute to 1 litre with water. Prepare a lOO-pg/ml solution by diluting 20 ml of this stock solution to 200 ml with water. Prepare the diluted solution fresh as required. Acetylacetone, 20% solution. Dilute 20 ml of acetylacetone and 30 ml of ethyl alcohol to 100 ml with water. This solution is stable for at least one week. Sodium 10 mg/ml solution. Dissolve 12.7 g of sodium chloride in water and dilute to 500 ml. Perchloric acid, 50% v/v. Sulphuric acid, 1% v/v. Hydrogen peroxide, 30%. Ammonium acetate solution, 50%. Procedures Calibration solutions. Add 10 ml each of 50% perchloric acid and lO-mg/ml sodium solution to each of nine 100-m]
standard flasks; then, from a burette, add to the first eight flasks 1, 2, 3, 5, 10, 15, 20 and 25 ml respectively, of the lOO-pgg/ml standard aluminium solution. The last flask contains the zero calibration solution. Dilute each solution to volume with water (Note 3). Iron, steel and nickel-chromium alloys. Transfer 0.1-l g of sample (Note 4) containing not more than 2 mg of aluminium to a 400-ml beaker, then cover the beaker and add 15 ml of concentrated hydrochloric acid and 5 ml of concentrated nitric acid. Heat until the sample has dissolved, then add 20 ml of concentrated perchloric acid (Note 5) and evaporate the solution to fumes of perchloric acid. Remove the cover and carefully evaporate the solution to 5-7 ml. Cool, add -50 ml of water and heat to dissolve the salts. Add about a quarter of a Whatman filter pulp tablet, macerate it with a stirring rod, then filter the solution (Whatman No. 40 paper) into a 400-m] beaker and transfer the residue quantitatively to the filter paper. Wash the paper and the residue three times with water, then three times with 1% sulphuric acid to ensure the complete removal of perchloric acid. Evaporate the filtrate to about 75 ml (Note 6). Carry a blank through the whole procedure. Transfer the paper containing the residue to a 30-ml platinum crucible, burn off the paper at a low temperature and ignite at 600-700”. Cool the crucible and add 1 ml each of 50% sulphuric acid and concentrated nitric and hydrofluoric acids. Heat gently to dissolve the residue, then evaporate the solution to dryness. Fuse the residue with 2 g of fused sodium hydrogen sulphate (Note 7), then cool the crucible and transfer it to a 250-ml beaker containing 4 50 ml of water. Cover the beaker and heat gently to dissolve the melt. Remove the crucible after washing it thoroughly with water (Note 8), then add the solution to the initial filtrate. Transfer the solution to a mercury-cathode cell, dilute to -200 ml with water and electrolyse for 75 min at -_ 10 A. Filter the electrolyte (Whatman No. 541 paper) into the 400-ml beaker that initially contained the solution. Wash the cathode cell and the paper thoroughly with water, then discard the paper. Add 5 ml of concentrated hydrochloric acid and 3 ml of 50”% sulphuric acid and evaporate the solution until 1.5-2 ml of sulphuric acid remain (Note 9). Cool the beaker, add 50 ml of water, heat gently until the solution is clear (Note lo), then cool it to room temperature. In succession, add 2 ml of 30% hydrogen peroxide, 5 ml of 20% acetylacetone solution and 10 ml of 50% ammonium acetate solution to the solution, then, using a pHmeter, adjust the pH of the solution to 6.5 f 0.1 with concentrated ammonia solution. Transfer the solution to a 125-ml separatory funnel, add 10 ml of chloroform, stopper and shake for 2 min. Allow several min for the layers to separate, then drain the chloroform phase into a 100-ml beaker. Extract the aqueous phase twice more by shaking it for 1 min each time with 5 ml of chloroform, then wash it by shaking it for -30 set with 5 ml of chloroform. Add 5 drops of concentrated hydrochloric acid to the combined extracts and evaporate to dryness in a hot water-bath. Add 1 ml of 50% perchloric acid and 2 ml of concentrated nitric acid, then cover the beaker and heat on a hot plate until fumes of perchloric acid are evolved (Note 11). Cool the solution to room temperature and add -3 ml of water. Transfer the blank solution to a lo-ml standard flask containing 1 ml of lo-mg/ml sodium solution and dilute to volume with water. Depending on the expected aluminium content, transfer the sample solution to a standard flask of appropriate size (lo-100 ml) and add sufficient lo-mg/ml sodium solution for 1 ml to be present for each 10 ml of final solution. If necessary, add sufficient 50% perchloric acid for 1 ml to be present for each 10 ml of final solution in excess of 10 ml (Note 12). and dilute the solution to volume with water. Measure the absorbances at 309.3 nm when the resulting
463
Determination of aluminium in iron, steel and alloys solutions are aspirated into a strongly reducing nitrous oxide-acetylene flame (Note 13). Calculate the aluminium contents (in mg) from the sample and blank absorbances and those obtained concurrently for calibration solutions that bracket the sample and blank concentrations. Correct the result obtained for the sample solution by subtracting that obtained for the blank solution. Ferrouanadium.Decompose 0.1-0.5 g of sample containing up to 2 mg of aluminium as described above, using 20 ml of 50% sulphuric acid instead of concentrated perchloric acid. Heat the solution until the evolution of oxides of nitrogen ceases, then remove the cover and evaporate to 5-7 ml. Cool, add 50 ml of water, heat to dissolve the salts, then proceed with the filtration of the solution (Note 14) the treatment of the residue, the mercury-cathode separation (Note l5), the extraction (Note 16) and the subsequent determination of aluminium as described above. Copper-baseand nickel-copper alloys.Transfer 0.1-l g of sample containing up to 2 mg of aluminium to a 400-ml beaker (Note 17). then cover the beaker and add 5 ml each of water and concentrated nitric acid and 20 ml of SOY0 sulphuric acid. Heat the solution until the evolution of oxides of nitrogen ceases, then remove the cover and evaporate to 5-7 ml. Cool, add 50 ml of water and, if necessary, heat to dissolve the salts. Transfer the solution to an electrolytic cell, dilute to 200 ml with water and proceed with the mercury-cathode separation (Note IS), the extraction and the subsequent determination of aluminium, as described above. Notes 1. A strongly reducing, non-luminous nitrous oxideacetylene flame is required to obtain the highest sensitivity for aluminillm. The height at which the beam from the hollow-cathode lamp passes through the flame is also extremely important. 1.3 Consequently, after all other instrumental parameters have been set, the acetylene and nitrous oxide flow-rates should be adjusted to give the maximum absorbance when a solution containing aluminium is aspirated into the flame. 2. Iron hastens the dissolution of the aluminium metal. 3. The calibration solutions should be prepared fresh every week. 4. More than 1 g of sample is not recommended because too much iron may remain in the solution after the mercury-cathode separation. The residual iron is co-extracted and, if a large amount is present, may result in an explosive reaction when the final solution is evaporated to fumes of perchloric acid. Too much iron can also cause a slight positive error in the aluminium result. 5. For iron and steel of low chromium content and high titanium, vanadium, molybdenum or tungsten content, add 20 ml of 50% sulphuric acid instead of perchloric acid and proceed as described for ferrovanadium. 6. If the acid-soluble aluminium content of the sample is required, proceed with the mercury-cathode separation, the extraction and the subsequent determination of aluminium as described. 7. One gram of a mixture consisting of 75% by weight of sodium carbonate and 25% by weight of fused boric acid can also be used for fusion of the insoluble material.4 However, the melt should be dissolved in water containing _ 2 ml of 50% sulphuric acid. 8. If the acid-insoluble aluminium content of the sample is required, it can be determined in the resulting solution. In this case, omit the mercury-cathode separation, add the volumes of hydrogen peroxide, acetylacetone and ammonium acetate solutions recommended in the subsequent part of the procedure, and proceed with the pH-adjustment, the extraction and the determination of aluminium as described. The total amount of aluminium present in the sample can then be determined by adding that found in the
acid-insoluble material to that found in the filtrate (Note 6). The separate determination of acid-insoluble aluminium is recommended if an appreciable amount of tungsten trioxide is present in the acid-insoluble material. In this case, add 5 ml of 30% hydrogen peroxide and the recommended volumes of acetylacetone and ammonium acetate solutions to the solution obtained after dissolving the melt, then add sufficient concentrated ammonia solution to dissolve the tungsten trioxide. Acidify the solution with 50% sulphuric acid, adjust the pH to 6.5 f 0.1 and proceed with the extraction and subsequent determination of aluminium. 9. This step volatilizes any mercury left in the solution. Sulphuric acid solutions of aluminium should not be evaporated to dryness. This results in the formation of an anhydrous aluminium sulphate that is virtually insoluble in water or in dilute acid and causes a low result for aluminiurn.’ 10. If the extraction cannot be completed the same day, allow the solution to stand overnight at this point. 11. The solution should not be allowed to evaporate to dryness. If the atomic-absorption determination cannot be completed the same day, allow the solution to stand overnight at this point. The final aluminium solution is not always stable on prolonged standing. 12. Additional 50% perchloric acid is not required if the final volume of the solution is to be 10 ml, because 1 ml is added to the residue remaining after the removal of chloroform. For final sample solution volumes of 25, 50 or 100 ml, add 1.5, 4 or 9 ml, respectively. 13. Between 5- and lO-fold scale expansion is recommended for the determination of aluminium. 14. It is not necessary to wash the paper and residue with 1% sulphuric acid as described in the procedure for iron and steel. 15. It is not necessary to add 3 ml of 50% sulphuric acid after the mercury-cathode separation. 16. More than 2 ml of 30% hydrogen peroxide-as recommended in the procedure for iron and steel-will be required for samples of high vanadium content. Approximately I ml is required to complex 30 mg of vanadium. Five millilitres should be added for a 0.5-g sample of ferrovanadium containing about 30% of vanadium. The solution should be yellow-not green-after the pH-adj stY ment. 17. If the sample contains silicon, use a 400-ml Teflon beaker and add 1 or 2 ml of concentrated hydrofluoric acid during the decomposition step. RESULTS
Atomic-absorption finish after separation of aluminium by extraction of its acetylacetone complex
Tests showed that, after the separation of aluminium by extraction of its acetylacetone complex into chloroform at pH 6.5 as described earlier,’ aluminktm can readily be determined by AAS in a nitrous oxideacetylene flame after evaporation of the extract and a small amount of concentrated hydrochloric acid to dryness and destruction of the organic material in the residue with nitric and perchloric acids as described in the proposed method. At least 0.5 ml of concentrated perchloric acid is required for this decomposition step. Consequently, a perchloric acid concentration of 5% by volume was chosen for the sample and calibration solutions because a final sample solution volume of at least 10 ml is necessary for the determination of small amounts of aluminium. To
464
ELSIE M. DONALDSON
suppress the ionization of aluminium, 1000 pg of sodium (as the chloride) per ml was also added to the sample and calibration solutions.’ EfSpct of diverse ions The electrolyte obtained after the separation of iron, chromium and other matrix elements from a dilute perchloric or sulphuric acid solution of an iron or steel sample by electrolysis with a mercury-cathode would contain the titanium, zirconium, vanadium and phosphorus (as phosphate) present in the sample, some of the manganese and tungsten and small amounts of residual iron, chromium, molybdenum and possibly nickel, cobalt and copper. It was shown previously’ that tungsten, niobium and tantalum can be kept in solution, and that molybdenum, titanium and vanadium can be prevented from co-extracting as acetylacetone complexes, by complexing them with hydrogen peroxide during the extraction of the aluminium-acetylacetone complex into chloroform at pH 6.5 from a 0.05M acetylacetone medium. Under these conditions, only iron(III), copper(I1) and beryllium are completely co-extracted. Cobalt(II), lead and chromium(III), which is produced in an acid medium containing hydrogen peroxide, are very slightly coextracted, and more than about 8 mg of phosphate causes a low result for aluminium because of the formation of aluminium phosphate. Tests, in which aluminium was extracted from a O.lM acetylacetone medium to increase the range of the proposed method, showed that up to at least 10 mg of iron(II1) or 5 mg of copper(I1) will not interfere in the extraction of up to at least 2 mg of aluminium. Furthermore, these quantities of iron and copper will not cause significant error in the aluminium result when the final solution is diluted to 10 ml. A larger amount of iron can cause a slight positive error ( - 3% at the lOO+g level) in the result. However, analysis of the final sample solutions derived from l-g samples of iron and steel showed that usually less than 5 mg of iron will remain in the electrolyte after the mercurycathode separation. Up to at least 10 mg of cobalt(II), lead, nickel and chromium(III), 20 mg of manganese(I1) and phosphate, and 5 mg of niobium and tantalum can be present during the extraction step without interfering in the extraction or subsequent determination of aluminium. In the presence of 2 ml of 30% w/v hydrogen peroxide, up to at least 60 mg of vanadium(V) and 20 mg of molybdenum(VI), tungsten(V1) and titanium(IV) will not interfere in the extraction of aluminium. Interference from up to at least 150 mg of vanadium can be avoided by adding 5 ml of 30% hydrogen peroxide before the pH-adjustment. Any insoluble tungsten trioxide present at this stage can be dissolved by adding concentrated ammonia solution. Subsequently, the solution should be acidified with dilute sulphuric acid followed by the adjustment of the pH to 6.5 + 0.1 with concentrated ammonia solution.
The effect of beryllium on the determination of aluminium by AAS was not investigated because it is not usually present in iron and steel. Although it was found previously’ that small amounts of zirconium can interfere in the extraction of aluminium by causing emulsification in the chloroform phase, up to at least 0.5 mg will not interfere in the proposed method. Applications To test the reliability of the proposed method, it was applied to the analysis of National Bureau of Standards (NBS) and British Chemical Standards (BCS) certified reference iron, steel and ferrovanadium samples and to a stainless steel to which known amounts of aluminium were added. It was also applied to NBS copper- and nickel-base alloys. The results of these analyses, which are the means of four or five AAS measurements, are given in Table 1. All the results shown are for individual samples. The results obtained previously for some of these reference materials by the spectrophotometric Pyrocatechol Violet method’ are also given in Table 1. Precision The precision at about the 0.0005 and the 0.002% levels of aluminium was tested by analysing one series of five samples of a BCS high-silicon steel for both acid-soluble and acid-insoluble aluminium. A second series of five samples was analysed for total aluminium. Although this reference material has not been certified for aluminium, results for acid-soluble, acidinsoluble and total aluminium have been reported by previous investigators. 4.5*‘o The results of these tests, which are the means of four or five AAS measurements on each sample solution, are given in Table 2.
DISCUSSION
Table 1 shows that the results obtained for the NBS and BCS iron, steel, ferrovanadium and non-ferrous alloys are in good agreement with the certified values and, where applicable, with the results obtained previously by the Pyrocatechol Violet method.’ The results obtained for the stainless-steel sample to which known amounts of aluminium were added, agree with the total calculated amount present. Table 2 shows that the relative standard deviation for acid-soluble aluminium at the 0.0004~~ level-and consequently also for total aluminium-is relatively high. This is partly due to the magnitude of the reagent blank, which is given below. At the 0.002% level, the precision of the results for acid-insoluble aluminium, for total aluminium determined by adding acid-insoluble and acid-soluble aluminium and for total aluminium determined in one step is comparable with that obtained by Shaw and Ottaway” by flameless AAS and with that obtained by Jenkins and Jones5 and other investigators using conventional
Determination of aluminium in iron, steel and alloys
465
E a~
o
e,¢
o., eq-
a~
E O
rz O
Z
¢.
g G
.E
o
~g
e-
llt
=
I I
1
2
g~ o
0
Z .E
Z
E
.o
E
E
._~ .E N
o
~,,,E
a
. .O.
.~_
Z
l
~
UN
~N
~NUo
E
•
~
~
N ~ ..~ ++
I--
.E~ E
±i
~.~.~
E~
~ m ZZZZZ
~'~ ZZ
Z
ZZmm~mZZZZZ
ZZ
466
ELSIEM. DONALDSON
AAS.4*6 The precision of the results obtained for total aluminium determined in one step is better than that obtained when acid-soluble and acid-insoluble aluminium are determined separately and added. However, this is expected from statistical considerations and because of the fewer manipulations required. Table 2 also shows that the results obtained for acid-soluble, acid-insoluble and total aluminium by the proposed method are in excellent agreement with those obtained by previous investigators. Although sodium hydrogen sulphate is recommended for fusion of the acid-insoluble material in methods for determining total aluminium in iron, steel and various other materials,ll~lZ Cobb et aL4 found that the use of this flux produced variable and low results for aluminium in several BCS mild steel samples. Consequently, they recommended a mixture of sodium carbonate and boric acid. The particular reference materials used in their tests were not immediately available. However, the results obtained for acid-insoluble and total aluminium in the BCS highsilicon steel by the proposed method (Table 2) show that sodium hydrogen sulphate is a suitable reagent for fusion purposes. This is also indicated in Table 1. The results obtained for NBS samples 19g, 55e and 1OOA by the proposed method involving the use of sodium hydrogen sulphate for fusion of the acidinsoluble material are in excellent agreement with those obtained (shown in brackets) when a mixture of sodium carbonate and fused boric acid was used for fusion purposes. The proposed method has some definite advantages over the recent AAS method6 mentioned earlier, which is also applicable to high-alloy steels, and which involves a preliminary separation of iron by methyl isobutyl ketone extraction of the chloro-complex. In this method, the sample, including the acidinsoluble material, is decomposed in a Teflon-lined (85-ml capacity cup) pressure bomb. Aluminium is ultimately separated from the bulk of the alloying constituents such as chromium and nickel by a double extraction of its acetylacetone complex into 50 ml of pure acetylacetone which, because of its low density, requires the use of two separatory funnels. The combined acetylacetone extracts are then concentrated to -7 ml by evaporation in an oil-bath at 150-160” and aluminium is determined by spraying the extract into the flame. This requires the use of calibration solutions prepared in a similar manner by extracting known amounts of aluminium into pure acetylacetone. Part of the manganese, lead, cobalt and nickel present in the sample, much of the niobium, titanium and vanadium, most of the copper, tin and zirconium, and all of the iron(M) remaining after the methyl isobutyl ketone extraction step will be coextracted with the aluminium. In the proposed method for iron and steel, 1 ml of acetylacetone is used and the aluminium complex is extracted into chloroform which, because of its high density, is convenient to use when multiple extractions are required.
Determination
of aluminium in iron, steel and alloys
The extract is easily evaporated to dryness in the presence of a small amount of concentrated hydrochloric acid, and aluminium is ultimately determined in an aqueous medium which facilitates the preparation of calibration solutions. After a mercury-cathode separation, and in the presence of hydrogen peroxide as a complexing agent, essentially only the residual iron and a small amount of chromium(II1) are co-extracted with the aluminium. As stated previously, at the l-g level, usually less than 5 mg of iron(II1) will remain in the electrolyte. Although the pressure-bomb decomposition step described in the Headridge and Sowerbutts method6 would be advantageous for the determination of total aluminium, this was not pursued in this work because of the low capacity (25 ml) of the Teflon cup in the available commercial acid-digestion bombs. Attempts to evaporate the solution obtained after dissolving 1 g of steel with aqua regia to low volume in a 25-ml Teflon cup resulted in considerable loss of sample by spray. The proposed method also has some advantages over the previous AAS methods involving the preliminary separation of iron by isobutyl acetate extraction of the chloro-complex.4*5 In these methods, _ 50-150 mg of iron remain in the aqueous phase after the extraction. This requires the addition of iron to the calibration solutions to eliminate error. In the proposed method, so little iron remains in the electrolyte after the mercury cathode separation that the addition of iron to the calibration solutions is not necessary. The results shown in Table 1 for the copper-base alloys were obtained after ultimately separating aluminium by acetylacetone extraction. However, tests showed that at the l-g copper level, less than 0.5 mg remains in the electrolyte after the mercury-cathode separation. Consequently, the extraction step would not usually be necessary for copper-base alloys, and probably not for most nickel-base alloys, unless appreciable titanium, tungsten or vanadium is present. However, the mercury-cathode separation should be carried out in a perchloric acid medium
TAL. 28/7A--0
467
instead of a sulphuric acid medium and the electrolyte should be evaporated to -0.5 ml if the final solution for the AAS determination is to be 10 ml. The proposed method is suitable for iron and steel containing as little as 0.0005% of aluminium and for copper- and nickel-base alloys containing w 0.0002~0 or more. However, the accuracy that can be obtained at these levels depends on the magnitude of the respective reagent blanks. In this work, the blank varied from c 12 to 20 pg of aluminium during the determination of total aluminium in iron and steel. Approximately two-thirds of the aluminium present in this blank resulted from the sodium hydrogen sulphate used for fusion of the acid-insoluble material. The blanks obtained after fusion with sodium carbonate and fused boric acid were of the same magnitude. Those obtained during the analysis of the non-ferrous materials varied from 4 5 to 8 pg of aluminium.
REFERENCES
1. M. D. Amos and P. E. Thomas, Anal. Chim. Acta, 1965, 32, 139. 2. P. Konig, K.-H. Schmitz and E. Thiemann, Z. Anal. Chem., 1969, 244,232.
3. J. Y. Marks and G. G. Welcher, Anal. Chem., 1970, 42, 1033. 4. W. D. Cobb, W. W. Foster and T. S. Harrison, Lab. Pratt., 1975, 24, 143.
5. R. H. Jenkins and C. P. Jones, Proc. Sot. Anal. Chem., 1972, 9, 266. 6. J. B. Headridge and A. Sowerbutts, Analyst, 1973, 98, 51.
I. E. M. Donaldson, Talanta, 1971, 18, 905. 8. 1. M. Kolthoff and P. J. Elving, Treatise on Analytical Chemistry, Part II, Vol. 4, p. 377. Interscience, New York, 1966. 9. W. Slavin, Atomic Absorption Spectroscopy, p. 79. Interscience, New York, 1968. 10. F. Shaw and J. M. Ottaway, Analyst, 1975, 100, 217. 11. Annual Book of ASTM Standards, Chemical Analysis of Metals; Sampling and Analysis
of Metal Bearing Ores,
Part 12, pp.-549 653, 655,. 692:American Society for Testina and Materials. Philadelnhia. 1979. 12. Z. St&k, P. Povondra and J. Doleial, CRC Crit. Reu. Anal. Chem., 1977, 6, 255.