292 Analysb, May, 1968, Vol. 93, Pp. 292-297 An Indirect Amplification Procedure for the Determination of Niobium by Atomic-absorption Spectroscopy BY G. F. KIRKBRIGHT, A. M. SMITH AND T. S. WEST (Chemisty Department, Imperial College, Londm, S. W.7) Niobium in the range 5 to 50pg is determined by an amplification procedure in which rnolybdoniobophosphoric acid is formed and extracted into butanol. The molybdophosphoric acid, which is also formed, is first selectively extracted away from the molybdoniobophosphoric acid into isobutyl acetate. The eleven rnolybdate ions associated with each niobium atom are determined by direct atomic-absorption spectroscopy in a nitrous oxide - acetylene flame at 313.2 nm. Of the twenty-eight other ions studied, only titanium causes appreciable interference, although its presence at the same concentration as.niobium can be tolerated. Large amounts of tantalum do not interfere. SEVERAL spectrophotometric methods have been reported for the determination of niobium, in which the coloured heteropoly complex formed between niobium , phosphate and molybdate in acidic solution is ~ s e d . ~ , ~ , ~ In these methods the molybdoniobophosphoric acid (MNPA) is reduced to “heteropoly blue,” after the selective decomposition of the excess of binary molybdophosphoric acid (MPA), by increasing the acidity of the solution, The methods reported are not sensitive, and strict adherence to the time of addition of acid and reducing agent, as well as measurement of the heteropoly-blue absorbance, must be observed.Babko and Shkaravskii4 have investigated the equilibria governing the formation of the molybdo- niobophosphoric acid, and its composition in aqueous solution. They report the formation, in an acid medium,. of a stable complex, in which the combining ratio of niobium - phosphorus - molybdenum is 1 : 1 : 11. The same authors also obsexved that, whereas molybdophosphoric acid is readily extracted into butyl acetate at pH 1, molybdoniobophosphoric acid is not extracted. ShkaravskiF has reported that tantalum does not form a ternary heteropoly complex with molybdenum a d phosphate. We have previously reported the sensitive indirect determination of silicon and phosphorus6 by atomic-absorption spectroscopy of the twelve molybdenum atoms associated with each atom of these elements in molybdosilicic and molybdophosphoric acids, after their selective extraction into an organic solvent.It seemed that a sensitive determination of niobium with high selectivity might similarly result from formation and extraction of the molybdoniobophosphoric acid, and the determination by atomic-absorption spectroscopy of the eleven molybdenum atoms associated with each niobium atom in the heteropoly complex. In the procedure reported here, MNPA is formed in 0.5 M hydrochloric acid by addition of molybdate and phosphate to the sample solution, and the excess of phosphomolybdic acid is extracted into isobutyl acetate. The MNPA in the residual aqueous phase is then extracted into butanol; the organic phase is washed with 0-5 M hydrochloric acid and the molybdenum present determined by atomic absorption in a nitrous oxide - acetylene flame at 313.2 nm.The high sensitivity of the method reported here for the determination of niobium results from the sensitivity obtainable for the determination of molybdenum in the nitrous oxide - acetylene flame ; the 1 1-fold amplification available from use of the molybdoniobophosphoric acid; the enhanced absorbance for molybdenum at 313-2 nm in butanol by virtue of efficient nebulisation of the solvent ; and the concentration of the molybdenum into a smaller volume of solution by extraction into butanol. The method is selective because few other elements form extractable heteropoly complexes under the conditions used. 0 SAC and the authors.KIRKBRIGHT, SMITH AND WEST 293 EXPERIMENTAL APPARATUS- Techtron AA4 flame spectrophotometer, fitted with a 5-cm nitrous oxide - acetylene burner and molybdenum high intensity hollow-cathode lamp.The instrument settings used for the determination of molybdenum were: slit width, 100 pm; wavelength, 313.2 nm; nitrous oxide pressure, 15 p.s.i. ; the acetylene flow-rate was adjusted to produce a flame with a “red-feather” of about 1 to 2 cm height. The burner height was adjusted so that as much as possible of the light from the hollow-cathode lamp passed through this red zone. REAGENTS- All reagents should be of analytical-reagent grade; we found that reagents conforming to the AnalaR specifications were suitable. Molybdate reagent solutio%-Dissolve 10.69 g of ammonium molybdate tetrahydrate, (NH,),Mo,0,,.4H20, in distilled water and dilute to exactly 1 litre.Phosphate reagent solzction-Dissolve 0.1098 g of potassium dihydrogen orthophosphate, KH2P0,, in distilled water and dilute to exactly 1 litre; 1 ml of solution contains 25 pg of phosphorus. Wash l i p i d for butanol phase-Saturate hydrochloric acid (0.5 M) with butanol, and store in polythene bottles. Isobutyl acetate and b.utamZ-Generd-purpose reagent grade. Niobium-Spectrographically standardised niobium powder (Johnson and Matthey Ltd.) . PREPARATION OF NIOBIUM STOCK SOLUTIONS- Silicon-free niobium stock solutions were prepared by the following procedure. Dissolve 0.1000 g of niobium powder in 10 ml of 40 per cent. hydrofluoric acid containing a few drops of concentrated nitric acid. Evaporate the solution almost to dryness in a platinum crucible.Dissolve the residue in 10 ml of warm 40 per cent. hydrofluoric acid. Transfer the solution to a polythene bottle and dilute to exactly 100 ml with distilled water. This solution contains 1 mg of niobium per ml. Transfer a 5-ml aliquot of this solution to a platinum crucible and evaporate to dryness. Dissolve the residue in 5 ml of 40 per cent. hydrofluoric acid, add a few drops of concentrated sulphuric acid, and evaporate the solution until no further evolution of white fumes occurs. Cool, re-dissolve the residue, then repeat the evaporation with hydro- fluoric and sulphuric acids. Make a final evaporation of the solution of the residue with 5 ml of 40 per cent. hydrofluoric acid without addition of sulphuric acid.Dissolve the residue in the minimum amount of hydrofluoric acid (about 2 ml) and dilute to exactly 100 ml with distilled water. The silicon-free niobium solution contains 50pg per ml of niobium. PREPARATION OF CALIBRATION GRAPH FOR NIOBIUM- To each of a series of six 100-ml separating funnels, add 10 ml of the molybdate reagent solution, 0, 0.2, 0.4, 0.6, 0.8 and 1.0 ml of the niobium solution (containing 50 pg of niobium per millilitre), 10 ml of the phosphate reagent and sufficient 6 M hydrochloric acid and water to make the final. solution 0.5 M with respect to hydrochloric acid. Mix the solutions, allow to stand for about 5 minutes, and proceed with each as follows. Extract the MPA into two 25-mI portions of isobutyl acetate by shaking each time for 1 minute and discarding the organic phase.Add 10 ml of butanol to the remaining aqueous phase and shake for 1 minute to extract the MNPA. Some free molybdate is extracted at this stage into the butanol, but it is readily removed by rapid successive shakings with three 10-ml aliquots of butanol- saturated 0.5 M hydrochloric acid. Spray the butanol phase directly into the nitrous oxide - acetylene flame and determine the molybdenum content by atomic-absorption spectroscopy. PREPARATION OF SAMPLE SOLUTIONS- Samples whose niobium content is to be determined should be dissolved in a mixture of hydrofluoric and nitric acids and treated further to remove silicon, as detailed in the instructions for the preparation of the niobium stock solution.294 KIRKBRIGHT, SMITH AND WEST: AN INDIRECT AMPLIFICATION [Analyst, Vol.93 RESULTS AND DISCUSSION FORMATION AND EXTRACTION OF MNPA- We have already reported the optimal conditions for the determination of molybdenum by atomic-absorption spectroscopy in a nitrous oxide - acetylene flame'; and the conditions were not, therefore, studied here. It is not possible to form MNPA in aqueous solution without the simultaneous formation of MPA. Consequently, throughout the studies undertaken to determine the optimal solution conditions for the formation and extraction of MNPA, the following general procedure was adopted to remove the MPA formed. Both MPA and MNPA were allowed to form in acidic solution, and then the MPA was extracted into two 25-ml aliquots of isobutyl acetate before the extraction of MNPA into 10ml of butanol.The molybdenum associated with the niobium in the MNPA was then determined by direct atomic-absorption spectroscopy in the washed butanol phase. Each solution was measured against a blank produced under identical conditions, but containing no niobium. EFFECT OF ACIDITY- The effect on absorbance of variation in the acidity of an initial aqueous solution, which was 2-7 x 10-2 M in molybdate, 3.6 x lo4 M in phosphate and contained 40 pg of niobium, is shown in Fig. 1. I I I I I I I 0 0.2 0.4 0 6 0.8 I .o I -2 Hydrochloric acid concentration (molarity) Fig. 1. Effect on absorbance for 40 pg of niobium of molarity of hydrochloric acid during formation of molyb- doniobophosphoric acid EFFECT OF MOLYBDATE AND PHOSPHATE CONCENTRATION- The effect on absorbance of variation of the phosphate concentration of the initial aqueous solution, which contained 30pg of niobium and was 2.7 x 1 0 - 2 ~ in molybdate and had an over-all acidity of 0.5 M, is shown in curve A of Fig.2. The effect of varying the molybdate concentration of the initial 3.6 x 10"' M aqueous phosphate solution at the same acidity, but containing 25 pg of niobium, is shown in curve B of Fig. 2. EXTRACTION OF MPA AND MNPA- When an over-all hydrochloric acid concentration of 0.5 M was used, the molybdate and phosphate concentrations in the initial aqueous phase were standardised at 2.7 x M and 3.6 x 1 0 4 ~ , respectively. Under these conditions it was shown, from determinations by atomic absorption of the molybdenum in the organic phase, that two extractions with 25-ml aliquots of isobutyl acetate quantitatively remove the excess of MPA from the aqueous phase.Radiotracer experiments with niobium-96 proved that no MNPA was extracted into isobutyl acetate under these conditions. After these isobutyl acetate extractions, treatment with a single 10-ml aliquot of butanol was shown to extract the MNPA quantitatively.May, 19681 PROCEDURE FOR THE DETERMINATION OF NIOBIUM 296 In our previous work on the determination of silicon by a similar amplification procedure, it was found that butanol extracts some of the excess of molybdate. Consequently, in all the present experiments, the butanol phase was washed with butanol saturated with dilute hydrochloric acid before determining its molybdenum content by atomic absorption.Curve C in Fig. 2 shows the effect of washing the final butanol phase with various concentrations of hydrochloric acid. From the results obtained it was decided that three rapid washes with 10-ml aliquots of butanol-saturated 0.5 M hydrochloric acid were sufficient. " A B C I I I I I 0 200 400 '600 800 loo0 pg of phosp )rus (A) 0 2 4 6 8 lox 1 0 , pg of molybdenum (8) 0 0.4 0.8 1.2 1.6 2.0M hydrochloric acid (C) Concentration Fig. 2. Curve A, effect seen on absorbance for 30 pg of niobium of variation in phosphate reagent concentration (solution 0-5 M in hydrochloric acid, 2.7 x 1 0 - e ~ in molybdate) : curve B, effect seen on absorbance for 25 pg of niobium of variation in molybdate reagent concentration (solution 0-5 M in hydrochloric acid, 3.6 x l o - 4 ~ in phosphate) ; curve C, effect seen on absorbance for 25 pg of niobium of variation in molarity of hydrochloric acid used to wash final butanol phase MOLYBDENUM-TO-NIOBIUM COMBINING RATIO IN MNPA- Babko and Shkaravskii4 have observed that the composition of the MNPA in solution depends on the acidity at its formation, and report that the molybdenum-to-niobium ratio is 11 : 1 at pH 0 to 1.0, 11 : 2 at pH 1.5 to 3.5 and 11 : 3 at pH 4-5 to 5.5.Thus the best amplification ratio is obtained by forming the complex at pH 0 to 1.0. An experiment was devised to establish whether the MNPA was formed and extracted stoicheiometrically as the 11 : 1 complex under the conditions used in our procedure. Different amounts of niobium were taken through the procedure, and the MNPA in the final butanol phase was decomposed and back-extracted into dilute ammonia solution and diluted to a known volume (25 ml) with distilled water. The molybdenum content of the aqueous solutions was determined by atomic-absorption spectroscopy in a nitrous oxide - acetylene flame.The solutions prepared for the calibration graph for molybdenum in aqueous solution were saturated with butanol and contained dilute ammonia solution. The results of these deter- minations gave a reproducible molybdenum-to-niobium ratio of 11.4 ( * 0-2) : 1. After correction for the small amounts of molybdenum not removed in the hydrochloric acid washing procedure, which constitutes the blank in the recommended procedure for niobium, a repro- ducible molybdenum-to-niobium ratio of 11.0 (k0.2) : 1 was obtained.Thus, under the recommended conditions, it is the 11: 1 MNPA that is formed and extracted quantitatively into butanol.296 KIRKBRIGHT, SMITH AND WEST : AN INDIRECT AMPLIFICATION [Analyst, Vol. 93 CALIBRATION GRAPH AND OPTIMUM CONCENTRATION RANGE- The calibration graph obtained when the recommended conditions are used is linear over the range 5 to 50pg of niobium. After subtraction of the reagent blank the graph passes through the origin, and the absorbances corresponding to these concentrations in the aqueous solution (0.22 to 2.2 p.p.m.) are 0.06 and 0-63. The blank, which results from the small amount of molybdate reagent not removed from the butanol phase by the washing procedure, is reproducible and may be subtracted from the absorbance produced by the standards taken through the procedure.This blank gives rise to an absorbance for molyb- denum in the flame of about 0.05 absorbance unit versus a butanol solvent blank. PRECISION- The replicate analysis of 11 samples, each with 23 ml of a solution containing 30 pg of niobium (1.30 p.p.m.), gave an average absorbance of 0.39, and the relative standard deviation was 0.009 absorbance unit or 2.4 per cent. EFFECT OF OTHER IONS- The effect of twenty-eight other ions on the absorbance produced in the determination of 30 pg of niobium by the recommended procedure has been investigated. An ion was con- sidered not to interfere when an error in the absorbance of less than twice the standard deviation of the determination of niobium alone (4.8 per cent.) was produced.The addition of 1 ml of 10-1 M solutions of the following ions in the determination of 30 pg of niobium caused no interference : aluminium, silver, bismuth, beryllium, calcium, cadmium, cobalt (11), copper, iron( 111) , nickel, magnesium, manganese( 11), antimony(V), zinc, sulphate, nitrate, fluoride and EDTA. In addition, the presence of 1 ml of 10-1 M solutions of the important refractory elements tantalum, tungsten and zirconium produced no interference when taken through the procedure. The addition of 1 ml of 10-1 M solutions of any of the above twenty-one ions corresponds to a range of excesses by weight over niobium from 30-fold to 700-fold. Of the elements that also form heteropoly acids under similar conditions, arsenic and germanium can be removed by volatilisation of their chlorides, and silicon is eliminated by volatilisation of fluorosilicic acid during the sample preparation procedure.Titanium, however, has been reported to form a molybdotitanophosphoric acids and, although the presence of large amounts causes serious interference, an equal amount of titanium (30pg) was found to be tolerable in the procedure. Attempts made to mask the interference of titanium were un- successful. Vanadium(V) and chromium(V1) are tolerable up to 3-fold and 45-fold excesses by weight, respectively. ACCURACY- samples. The results of these analyses are shown in Table I. Chemical analyses were performed for niobium in prepared solutions treated as unknown TABLE I DETERMINATION OF NIOBIUM IN PREPARED MIXTURES Niobium added to initial aqueous solution, pg per 25 ml 18 30 30 30 30 6 60 10 5 9 40 30 21 21 Other ions present, Pg Zr (3680) Cr (IV) ( 1350) W(V1) (1840) Co (2945) : Ni (2935) Cu (3175) : Zn (3270) Fe(II1) (2800) Ta (300) Zr (736) Ta (600) Ta (30,000) W(V1) (1840) - - - Niobium found, PQ 18 30 30.6 29.5; 30 6 49 10.3 6 8.9 40-2 30 24* 21 * Results high because of traces of niobium in the tantalum used.May, 19681 PROCEDURE FOR THE DETERMINATION OF NIOBIUM 297 CONCLUSION The method reported here for the indirect determination of niobium by atomic-absorption spectroscopy is much more sensitive than the direct method for niobium in the nitrous oxide - acetylene flame (1 per cent.absorption is given by 24 p.p.m. of niobium by the direct method, whereas by this method I per cent.absorption results with an initial aqueous niobium solution of 0-015 p.p.m.). Because of the selectivity with which molybdoniobophosphoric acid can be formed and extracted, the selectivity of the procedure compared with the direct method is good. Extraction of MNPA into butanol achieves a separation of the niobium from other major element components of samples. In the direct method, the presence of large amounts of other elements frequently suppresses the niobium absorbance because of the formation of oxide in the flame, so that solvent extraction may be required. Thus the use of a solvent- extraction procedure in this indirect method does not constitute a disadvantage compared with the direct atomic-absorption method. The method reported is precise and, after the dissolution of the sample, almost as rapid as the direct procedure. We are grateful to the Ministry of Technology for support of this work, and to the Courtauld Research Foundation for a grant for the purchase of the atomic-absorption equipment used in these studies. 1. 2. 3. 4. 5. 6. 7. 8. Davydov, A. A., Vaisberg, 2. M., and Burkur, L. E., Zav. Lab., 1947, 13, 1038. Norwitz, G., and Codell, M., Analyt. Chern., 1954, 26, 1238. Veitsman, R. M., Zav. Lab., 1954, 25, 552. Babko, A. K., and Shkaravskii, Y. F., Russ. J . Inorg. Chem., 1962, 7 , 809. Shkaravskii, Y. F., Ibid., 1963, 8, 1399. Kirkbright, G. F., Smith, A. M., and West, T. S., Analyst, 1967, 92, 411. , I , Ibid., 1966, 91, 700. Babko, A. K., and Shkaravskii, Y. F., Russ. J . Inorg. Chern., 1961, 6, 1068. --- Received November 29th, 1967