Artalyst, March, 1968, Vol. 93, $@. 131-141 131 Spectrophotometric Determination of Niobium in Zirconium, Titanium and Other Metals with 4-(2=Pyridylazo)resorcinol BY D. F. WOOD AND J. T. JONES (Imperial Metal Industries Limited, Kynoch Works, Witton, Birmingham 6) A reaction based on the formation of a purple-coloured complex between niobium and 4- (2-pyridy1azo)resorcinol (PAR) in a tartrate - ethylenediamine- tetra-acetate (EDTA) solution has been successfully applied to the direct spectrophotometric determination of niobium, in the range 50 p.p.m. to about 1 per cent., in zirconium, zirconium alloys, hafnium, tungsten and molybdenum. By modification, this direct procedure can be used for the determination of down to 200 p.p.m. of niobium in titanium and its alloys. Following an extractive concentration, amounts of niobium down to about 10 p.p.m. can be determined in titanium, zirconium, and their alloys, and in hafnium, molybdenum and iron.For tungsten (and molybdenum), a preliminary collection of niobium on zirconium hydroxide is recommended. The effects of common alloying elements, and likely impurities, have been investigated. The absorption of the tantalum complex is only about one eighth of that of the niobium - PAR complex. The absorbance is measured at 550mp. THE level of niobium in zirconium, zirconium alloys and hafnium is usually below 100 p.p.m. Other analytical requirements involve the determination of niobium at levels below about 500 p.p.m. in titanium, titanium alloys and certain other metals, and these factors necessitate the provision of reliable analytical procedures for determining small amounts of niobium in all of these materials.The formation of coloured compounds of niobium with selective chromogenic reagents is generally accepted as the most satisfactory basis for determining small amounts of niobium. Reagents widely used for this purpose include hydrogen peroxide,l pyrogallol,2 8-hydroxy- quinoline3 and thiocyanate,4 but none of these reagents is suitable for the direct determination of niobium below about 100 p.p.m., because of inadequate sensitivity, poor selectivity, or instability of the complex. In recent years increasing attention has been given to the spectrophotometric deter- mination of small amounts of niobium, and several new reagents have been proposed; these include xylenol orange,5 tribromopyrogallol,6 4-(2-pyridylazo)resorcinol (PAR) ,7#* 99 and bromopyrogallol red.1° From published results it appeared that PAR and bromopyrogallol red in the presence of ethylenediaminetetra-acetate (EDTA), which was used to mask interfering ions, offered advantages in respect of selectivity and sensitivity, and these two reagents were selected for further examinat ion.Initial tests indicated that zirconium and titanium also formed coloured complexes with bromopyrogallol red, and the interference of 100mg of either of these metals could not readily be prevented. Preliminary tests with PAR, however, were more promising, and a more detailed examination of this reagent was made. 0 SAC and the authors.132 APPARATUS- A Unicam SP600 was used in the experimental work, and absorbance measurements were made at 550 mp (4-cm cells).All pH adjustments were made with a direct-reading pH meter. WOOD AND JONES : SPECTROPHOTOMETRIC DETERMINATION OF [Artaiyst, vol. 93 EXPERIMENTAL PREPARATION OF CALIBRATION GRAPH- A standard solution of niobium was prepared from 0.1 g of high-purity niobium, which was fused in potassium hydrogen sulphate. The cooled melt was dissolved in 50 ml of 20 per cent. tartaric acid solution, and the solution was diluted to 500ml in a calibrated flask. To a 25-ml aliquot of this solution were added 50 ml of 20 per cent. tartaric acid solution, and the solution was diluted to 500 ml so that 1 ml of solution contained 10 pg of niobium. Aliquots of this standard solution ranging from 1-0 to 5.0ml were transferred into a series of 100-ml beakers; each solution, and a blank (water only), was diluted to about 10 ml.The characteristic colour was developed by using conditions similar to those recommended by Belcher, Ramakrishna and West,? vix., to each solution, 10.0ml of 1 per cent. EDTA solution were added, the solution was adjusted to pH 6-0 with dilute ammonia solution (1 + 2), then 10-0 ml of 0.03 per cent. PAR solution and 5 ml of the ammonium acetate buffer solution7 were added. All of the solutions were allowed to stand for 1 hour, then transferred into separate 100-ml calibrated flasks and diluted to the mark. Absorbances were measured against the blank solution. After the absorbance of the reagent blank (0.035) had been deducted, the graph (absorbance against niobium) was a straight line that passed through the origin; an absorbance value of 0.55 corresponded to 50 pg of niobium.A further series of tests showed that addition of the acetate buffer solution, as recommended by Belcher, Ramakrishna and West,7 was unnecessary, provided that the pH of the test solution was adjusted to 6.0, immediately before, and after, addition of the PAR solution. Tests were made next in the presence of zirconium. Aliquots of the standard niobium solution (1 ml of solution = 10 pg of niobium) ranging from 1.0 to 5-0 ml were added to 0.1-g samples of high-purity zirconium; an additional sample of zirconium was used as a “blank.” The zirconium was dissolved in 10 ml of dilute sulphuric acid (1 + 49), 1.0 ml of fluoroboric acid solution and 2.0 ml of 50 per cent.tartaric acid solution; the presence of tartaric acid prevents the subsequent hydrolysis of zirconium salts when the niobium - PAR complex is developed. The solutions were cooled, the niobium - PAR complex was developed in the absence of the buffer solution, as before, and absorbances were measured. These were much higher than those obtained in the absence of zirconium, and this was attributed to the formation of a zirconium - PAR complex. In subsequent tests, the interference by zirconium was overcome by increasing the amount of EDTA. For example, with 10.0ml of 5 per cent. EDTA solution, the graph of absorbance (corrected for a blank of 0.04) against niobium was a straight line that passed through the origin; an absorbance value of 0.58 corresponded to 50pg of niobium.In the latter tests, a period of 2 hours was allowed for the colour to develop before the absorbances were measured, because it was observed by the previous authors7 that an increase in the amount of EDTA increases the time required for the reaction to reach equilibrium. The graph obtained was applicable to the determination of niobium in the range 50 to 500 p.p.m. in zirconium. Further tests showed that this upper niobium limit could be extended up to at least 1 per cent. by taking a suitable aliquot of the test solution and main- taining the volume (2.0 ml) of 50 per cent. tartaric acid solution in the final solution in which the colour is developed. CONCENTRATIONS OF THE REAGENTS- PAR-In tests with 50 pg of niobium and 1O*Oml of 0.03, 0-04 and 0.05 per cent.solutions of PAR, absorbance values ( r n i r t ~ s blank) were 0.58, 0.64 and 0.69, respectively; corresponding reagent blank values were 0.03, 0.05 and 0.075. An explanation of this progressive increase in absorbance with increase in concentration of the reagent is that the niobium - PAR complex is partially dissociated, and the degree of dissociation decreases as the amount of PAR is increased. Because the absorbance of the reagent blank alsoMarch, 19681 NIOBIUM IN ZIRCONIUM, TITANIUM AND OTHER METALS 133 increases as the amount of PAR is increased, a compromise was made to obtain a reasonably high sensitivity, together with an acceptable (low) blank value, and in all of the subsequent tests 10.0 ml of 0.04 per cent.PAR solution were used. When PAR solutions prepared from different batches of reagent were used, widely different sensitivities were obtained. The reagent used in the previous tests, in which an absorbance value of 0.64 was obtained, was supplied by British Drug Houses Ltd. Absorbance values obtained for 50 pg of niobium with PAR from two other sources were 0.55 and 0.74. Only a limited amount of the most sensitive of these reagents (supplied by Professor T. S. West) was available, and all of the subsequent tests were made with the British Drug Houses' reagent. EDTA-The calculated amount of EDTA required to form a complex with 0.1 g of zirconium is about 0.41 g ; 10.0 ml of 5 per cent. EDTA solution, as used in the previous tests, are, therefore, sufficient to provide a reasonable excess and for traces of interfering ions.Decrease in the concentration of EDTA below the theoretical amount required to react with 0.1 g of zirconium caused an increase in absorbance, as unchelated zirconium species react with PAR to form a strongly coloured complex. When the concentration of EDTA was increased above the equivalent of 10.0ml of a 5 per cent. solution, absorbances pro- gressively decreased; in the presence of 2 g of EDTA, an absorbance value of 060 was obtained for 50 pg of niobium, compared with 0.64 in the presence of 0.5 g of EDTA. This effect is probably caused by an increase in competition of the EDTA for the niobium ions. In subsequent tests 10.0 ml of 5 per cent. EDTA solution were used.Sulphuric acid a.nd$ztoroboric acid solutions-Satisfactory dissolution of 0-1 g of zirconium is effected with 1 ml of fluoroboric acid solution and 10 ml of dilute sulphuric acid (1 + 49). Variation in the amount of dilute sulphuric acid (1 + 49) used from 8 to 12 ml had no significant effect on the final absorbance. Increase in the volume of fluoroboric acid solution from 1 to 2ml had no significant effect on the absorbance, but when the volume of this reagent exceeded 2 ml, absorbances progressively decreased; with 50 pg of niobium, an absorbance value of 0.55 was obtained in the presence of 5 ml of fluoroboric acid solution, compared with 0.64 in the presence of 1 to 2 ml of the reagent. Tests on solutions containing no zirconium and variable amounts of fluoroboric acid, with other conditions the same as those used in the presence of zirconium, showed that an increase in the amount of fluoroboric acid solution of up to 1 ml produced a significant increase in absorbance.In the absence of fluoroboric acid, an absorbance value of only 0.40 was obtained for a solution containing 50 pg of niobium, compared with 0.66 in the presence of 1 nil of fluoroboric acid solution. The niobium - PAR complex has been shown to be anionic,' and it is likely that it contains tartrate. It is possible, therefore, that this increase in sensitivity with increase in fluoroboric acid is caused by replacement of tartrate in the complex by fluoroborate (or fluoride) ions, thus producing a complex that is more strongly co-ordinated. Tartaric acid-Tartaric acid is added to prevent the hydrolysis of zirconium and niobium salts. Tests showed that the minimum amount of tartaric acid necessary to prevent this hydrolysis completely (at pH 6) is about 0.5 g; a reasonable safety margin is provided in the presence of 2-0 ml of a 50 per cent.solution of the reagent. Increase in the amount of tartaric acid to 5 g caused a decrease in absorbance of about 30 per cent.; this is probably caused by increased competition of tartrate for the niobium ions. OTHER VARIABLES- PH-By using the best conditions established in the previous tests, absorbance was shown to be independent of pH over the range 6.0 to 6.4. In subsequent tests, the pH was adjusted to 6-2 immediately before, and after, addition of the PAR solution.Tests showed that if the PAR is added before adjusting the pH, low and erratic results are obtained. Temperature and colour development time-Tests on typical solutions showed that at 15" C the colour developed fairly rapidly during the first 15 minutes, but more slowly during the next 90 minutes; thereafter, it was constant for at least 3 hours. Similar tests at 20" and 30" C followed the same pattern, except that during the first 15 minutes the rate at which the colour developed increased with temperature. The final absorbance, measured after 2 hours, was the same in each series of tests. In subsequent tests, the colour was allowed to develop for 2 hours at 20" C before the absorbance was measured.134 [Autdyst, vol. 93 Other metals-Tests showed that the procedure used in the presence of zirconium was equally applicable to the determination of niobium in the presence of 0-1 g of hafnium.Satisfactory results were also obtained in the determination of 10 to 50pg of niobium in the presence of 0.1 g of tungsten. With 0.1 g of molybdenum, the absorbance was suppressed slightly, but the effect was not significant; larger amounts of molybdenum, however, could not be tolerated. Details for preparing solutions of these metals to enable the proposed method to be extended are given in Method I. Based on a 0.1-g sample, no significant interference was introduced by the presence of up to 25 per cent. of tin, 10 per cent. of aluminium, 5 per cent. of titanium, 1 per cent. of copper, 0.5 per cent. of iron, chromium, manganese or nickel, or 0.1 per cent.of vanadium. Larger amounts of iron up to about 5 per cent. could be tolerated by using 10 ml of a 5 per cent. solution of the disodium salt of 1,2-diaminocyclohexanetetra-acetate (CDTA) instead of the EDTA solution, but CDTA was less effective than EDTA in suppressing the inter- ference of titanium. The effect of CDTA in the presence of other interfering elements was not examined, and in subsequent tests EDTA was used as the complexing agent. When present in excess of the limiting concentrations specified above, copper interfered by forming a coloured complex with EDTA; iron, nickel and vanadium interfered by forming coloured complexes with PAR, and manganese suppressed development of the niobium - PAR complex. The effect of chromium and titanium was 2-fold; these metals reacted with PAR to form coloured complexes and also suppressed development of the niobium - PAR complex.Under conditions similar to those established for determining niobium in zirconium, tantalum also reacts to form a coloured complex. The molar extinction coefficient of the tantalum - PAR complex is about 8000, whereas that of the niobium - PAR complex is about 30,000 but, because of their relative atomic weights, absorbance of the tantalum complex is about one eighth of that of the niobium complex. For most practical purposes, up to about 50 p.p.m. of tantalum can be tolerated in the determination of 50 to about 200 p.p.m. of niobium; proportionately more can be tolerated if the niobium content of the sample exceeds 200 p.p.m.The tantalum content of zirconium, hafnium, tungsten and moly- bdenum is usually less than 20 p.p.m., hence the effect of tantalum in the determination of niobium in these metals is not significant. Belcher, Ramakrishna and West7 have shown that the interference of tantalum can be minimised by increasing the tartrate concentration, and subsequent tests in the present investigation confirm this observation. According to Elinson, Pobedina and Rezova,B the interference of tantalum is eliminated by developing the niobium - PAR complex in the presence of a large excess of tartrate at pH 4-5 to 4.8. Under these conditions, however, the sensitivity of the reaction is considerably reduced, and their method is restricted to niobium contents above 0-1 per cent. WOOD AND JONES : SPECTROPHOTOMETRIC DETERMINATION OF EXTENSION OF THE PAR PROCEDURE TO TITANIUM AND TITANIUM ALLOYS- Earlier tests had shown that titanium in amounts above 5 per cent., based on a 0.1-g sample, interferes in the direct determination of niobium, and tests were designed to establish, more precisely, the extent of this interference, and, if possible, provide a suitable modification that would overcome the interference.The procedure developed so far was applied to solutions containing (a) 0.1 g of titanium and (b) 0.1 g of titanium and 50 pg of niobium. Absorbance values obtained in these tests were (a) 0.33 and (b) 0.38; corresponding values obtained in similar tests in which zirconium replaced titanium were 0.05 and 0.69, respectively. These absorbance values show the extent to which titanium forms a coloured complex with the PAR and to which it suppresses development of the niobium - PAR complex.The effects of titanium were shown to be considerably reduced when the amounts of EDTA and tartaric acid were increased to 1 and 5 g , respectively; solutions similar to (a) and (b) containing these increased amounts of EDTA and tartaric acid gave absorbance values of (a) 0.04 and (b) 0.42. In further tests under similar conditions, the graph (absorbance against niobium) over the range 20 to 1OOpg was a straight line that passed through the origin. However, because of the decrease in sensitivity caused by the increased amounts of these two reagents, the direct procedure (Method I) is only suitable for determining niobium in amounts above 200 p.p.m.March, 19681 NIOBIUM IN ZIRCONIUM, TITANIUM AND OTHER METALS 135 Tests showed that by increasing the amount of tartaric acid to 5 g, interferences from tantalum and vanadium were reduced, and up to 500 p.p.m.of tantalum, or 5 per cent. of vanadium, had no significant effect on the determination of niobium in amounts above 200 p.p.m. The effects produced by increasing the amount of tartaric acid or EDTA, or both, in the presence of other interfering metals were not examined. To make the fullest use of the inherent sensitivity of the proposed method for deter- mining lower levels of niobium in titanium-based samples, a method of separation, based on extraction of the niobium from a hydrofluoric acid - sulphuric acid - ammonium fluoride solution into isobutyl methyl ketonell was investigated.This extraction procedure was applied to solutions containing 0.1 g of titanium and amounts of niobium ranging from 10 to 50 pg; a solution containing 0.1 g of titanium was used as a blank. The separated organic phase was back-extracted with a solution of hydrogen peroxide, and niobium was subsequently determined with PAR, as in the direct procedure for the determination of niobium in zirconium. Absorbance values obtained in these tests, compared with those obtained by a direct application of the PAR method to solutions that contained niobium alone, showed that recoveries of niobium were between 94 and 99 per cent. The absorbance value of the blank solution (0.046) indicates that no significant amount of titanium is extracted with the niobium.Recoveries of 10 to 50 pg of niobium from solutions containing 0.5 g of titanium were equally satisfactory. This extraction procedure is, therefore, suitable for determining down to about 10 p.p.m. of niobium in titanium, and tests on solutions containing 0.5 g of titanium in the presence of certain other metals showed that the extraction also overcomes the interference of at least 5 per cent. of iron, copper, vanadium, chromium, manganese or nickel in the determination of between 10 and 50pg of niobium. As tantalum is also extracted with isobutyl methyl ketone, the extraction stage does not overcome the interference by tantalum. In further tests incorporating the extraction procedure, complete recoveries of 10 to 50 p g of niobium were obtained from 0-5 g of iron.Similar tests on solutions containing 10 to 50 pg of niobium in the presence of 0.5 g of zirconium or hafnium showed that the extraction enabled the determination of niobium in these metals to be extended down to about 10 p.p.m. In tests on solutions containing 10 to 50 pg of niobium and 0.5 g of molybdenum, the separation of molybdenum from niobium was not complete, but the amount of molybdenum extracted with the niobium did not significantly affect development of the niobium - PAR complex, and recoveries of niobium down to about 10 p.p.m. were satisfactory. In similar tests on solutions containing 0-5g of tungsten, a large proportion of the tungsten was extracted with the niobium, and difficulties were introduced by the precipitation of hydrated tungsten trioxide at later stages in the procedure, but recoveries from solutions containing only 10 mg of tungsten were satisfactory. In subsequent tests, the niobium was precipitated from an ammoniacal solution in the presence of added zirconium, which provided a zirconium hydroxide carrier; the precipitate was filtered off, calcined and then fused with sodium hydrogen sulphate before applying the niobium- PAR reaction.These tests showed that 10 to 50pg of niobium could be separated from at least 1 g of tungsten. This procedure was shown to be equally applicable to solutions containing up to at least 1 g of molybdenum; it is preferable to the extraction procedure, because a complete separation of niobium can be achieved from a relatively large amount of molybdenum.The extraction procedure, however, has the advantage that it is less susceptible to interference from certain alloying const it uent s. METHOD I This method is suitable for the direct determination of from 50 p.p.m. up to about 1 per cent. of niobium in zirconium, zirconium alloys (ZR20 and ZR30), hafnium, tungsten and molybdenum. It can also be applied to titanium if the niobium content is more than about 200 p,p.m. (Note 1). REAGENTS- 10" C) add, in small amounts, 130 g of boric acid. Store in a polythene bottle. Fluoroboric acid solution-To 280 ml of 40 per cent. hydrofluoric acid (maintained at136 WOOD AND JONES : SPECTROPHOTOMETRIC DETERMINATION OF [Analyst, Vol. 93 Standard niobium solution-Transfer 0.1 g of high-purity niobium (powder or millings) into a 100-ml conical flask, add 5 g of potassium hydrogen sulphate, heat the flask over a Meker burner until the metal has dissolved, then cool. Add 50 ml of 20 per cent.tartaric acid solution, warm gently until the melt has dissolved, then cool. Transfer the solution into a 500-ml calibrated flask and dilute to the mark. Transfer 25 ml of this solution into a 500-ml calibrated flask, add 50 ml of 20 per cent. tartaric acid solution, and dilute to the mark. 1 ml of solution = 10 pg of niobium. EDTA solution, 5 per cent.-Dissolve 50 g of the disodium dihydrate salt of ethylene- PAR solution, 0.04 per cent.-Dissolve 0-4 g of the disodium salt of 4-(2-pyridylazo)- diaminetetra-acetic acid (EDTA) in water, and dilute the solution to 1 litre.resorcinol (PAR) in water, and dilute the solution to 1 litre. PROCEDURE FOR NIOBIUM CONTENTS 50 TO 500 P.P.M. (NOTE 2)- Determine a reagent blank with each batch of samples and apply the procedure to two reference solutions, each containing 2.5 ml of the standard niobium solution (1 ml of solution = 10 pg of niobium). In the preparation of the reference solutions for use in the analysis of zirconium alloys, tungsten, molybdenum or titanium, add the standard niobium solution after the fuming stage with sulphuric acid, otherwise the tartaric acid in the standard niobium solution decomposes during fuming with the sulphuric acid and discolours the solution. PREPARATION OF SAMPLE SOLUTION- Zirconium and hafnium-Transfer a 0.1-g sample into a 50-nil beaker, add 2-Om1 of 50 per cent.tartaric acid solution, 10 ml of dilute sulphuric acid (1 + 49) and 1.0 ml of fluoroboric acid solution. Heat gently until the sample has dissolved, then cool the solution. Zirconium aZZoys (ZR20 and ZR3O)-Transfer a 0.1-g sample into a small platinum dish, add about 5ml of water, then 40 per cent. hydrofluoric acid, dropwise, until the sample has dissolved. Add a few drops of nitric acid (spgr. 1.42) (to dissolve precipitated copper or tin), cool and add 1 ml of sulphuric acid (sp.gr. 1.84). Evaporate the solution to fumes of sulphuric acid, fume for about 3 minutes, then cool. Add 2.0 ml of 50 per cent. tartaric acid solution, 5 ml of water, then transfer the solution into a 50-ml beaker with a minimum amount of water and add 1.0 ml of fluoroboric acid solution.Tungsten and molybdenum-Transfer a 0.1-g sample into a small platinum dish, add 2 ml of nitric acid (sp.gr. 1-42), dropwise, then 40 per cent. hydrofluoric acid, dropwise, until the sample has dissolved, and cool the solution. Add 1 ml of sulphuric acid (sp.gr. 1-84), evaporate to fumes of sulphuric acid, fume almost to dryness (about 10 minutes), then cool. Add 2 g of sodium hydrogen sulphate, heat gently over a Meker burner until a clear melt is obtained, then cool. Add 2-0 ml of 50 per cent. tartaric acid solution and about 5 ml of water; heat gently to dissolve any precipitated tungstic or molybdic acid, then cool. Transfer the solution into a 50-ml beaker with a minimum amount of water and add 1*0ml of fluoroboric acid solution.Titanium-See Note 1. COLOUR DEVELOPMENT- Proceed with the reagent blank, reference and sample solutions as follows. Add 10.0 ml of the 5 per cent. EDTA solution, then dilute ammonia solution (1 + l), dropwise, until the pH of the solution is 6.2 Add 10.0 ml of the 0.04 per cent. PAR solution. If necessary, adjust the pH to 6-2 with dilute ammonia solution (1 + 4) or dilute sulphuric acid (1 + 9) as required, and allow the solution to stand for 2 hours. Transfer the solution into a 100-ml calibrated flask, dilute to the mark, mix, then measure the absorbance at a wavelength of 550mp, with 4-cm cells. Deduct the blank value, and calculate the niobium content of the sample froin the niobium recoveries obtained on the reference solutions. 0.1 (use a pH meter).March, 19681 NIOBIUM I N ZIRCONIUM, TITANIUM AND OTHER METALS 137 NOTES- 1.Based on a 0.1-g sample, no significant interference is caused by the presence of up to 25 per cent. of tin, 10 per cent. of aluminium, 5 per cent. of titanium, 1 per cent. of copper, 0.5 per cent. of iron, chromium, manganese or nickel, 0.1 per cent. of vanadium or 50 p.p.m. of tantalum. For samples that contain alloying levels of copper or tin, or both, use the dissolution procedure recommended for zirconium alloys. The materials for which the method is recommended normally contain less than 20 p.p.m. of tantalum, but up to 500 p.p.m. of tantalum can be tolerated by increasing the volume of 50 per cent. tartaric acid solution to 10.0 ml; a similar amount must be added to the blank and reference solutions, because tartaric acid lowers the sensitivity of the niobium - PAR reaction.This direct procedure can be applied to the determination of 200 to 1000 p.p.m. of niobium in titanium (0.1 g), provided that the sample is dissolved as described for zirconium alloys, the tartaric acid solution (50 per cent.) is increased to 10.0 ml,* and 1 g of EDTA is added (the EDTA will not dissolve completely until the pH of the solution is adjusted to about 6.2). Higher concentrations of niobium in titanium, up to about 1 per cent., can be determined in an aliquot of the sample solution (containing not more than 100 pg of niobium) by adjusting the tartaric acid concentration to the equivalent of 10.0 ml of a 50 per cent. solution before developing the coloured complex.A blank, and reference solutions containing 5-0 ml of the standard niobium solution (1 ml of solution = 10 pg of niobium), must be examined in the same way as the sample solution; in the preparation of the reference solutions, add the standard niobium solution after the fuming stage with sulphuric acid. For determining lower levels of niobium in titanium, see Method 11. 2. For niobium contents below 50 p.p.m., use a larger weight of sample, and make a preliminary separation of niobium (see Methods I1 and 111). For niobium contents from 500 p.p.m. up to about 1 per cent. in zirconium, zirconium alloys and hafnium, use a 0.1-g sample and prepare the solutions as described under Zirconium alloys,? but add 4.0 ml of 50 per cent. tartaric acid solution. Add 5 ml of water, dilute the solution to 100 ml in a calibrated flask, and transfer an aliquot not exceeding 50 ml and containing not more than 50 pg of niobium into a 100-ml beaker.If necessary, add more of the 50 per cent. tartaric acid solution to bring the total amount present to 2.0 ml, add 1.0 ml of fluoroboric acid solution, and continue as described under Colour development. For niobium contents from 500 p.p.m. up to about 1 per cent. in tungsten and molybdenum, prepare solutions as described for these metals, but add 4.0 ml of 50 per cent. tartaric acid solution. Heat gently to dissolve any tungstic or molybdic acid, then cool. Transfer the solution into a 100-ml calibrated flask, dilute to the mark, and continue as described above in this Note for determining from 500 p.p.m.to 1 per cent. of niobium in zirconium. METHOD I1 This extraction procedure enables the PAR method to be extended down to about 10 p.p.m. of niobium in titanium, zirconium, hafnium, iron and molybdenum; it also enables niobium, up to about 500 p.p.m., to be determined in samples that contain alloying metals in excess of the limiting concentrations specified in Method I. SPECIAL REAGENTS- (See Notes 3 and 4.) In addition to those described for Method I. Isobuty2 methyl ketone (eqztiZibrated)-To 40ml of 40 per cent. hydrofluoric acid con- tained in a 250-ml polythene beaker add, cautiously, 33.5 ml of sulphuric acid (sp.gr. 1-84), then cool. Add 8-Og of ammonium fluoride, stir until the reagent has dissolved, then cool and dilute the solution to 100ml.Transfer the solution into a 500-ml polythene bottle, add 300ml of isobutyl methyl ketone, shake the mixture for 2 minutes, then allow the organic and aqueous phases to separate. Use 25-ml aliquots of the upper organic layer, as required, in the following procedure. PROCEDURE FOR NIOBIUM CONTENTS 10 TO 100 P.P.M. (NOTE 4)- Determine a reagent blank with each batch of samples and apply the procedure to two reference solutions, each containing 2.5 ml of the standard niobium solution (1 ml of soh- tion -= 1Opg of niobium). In the preparation of the reference solutions, add the standard niobium solution after the fuming stage with sulphuric acid. Transfer a 0.5-g sample into a small platinum dish, add about 5 ml of water, then 40 per cent. hydrofluoric acid (Note 5), dropwise, until the sample has dissolved.Add about 3 ml of nitric acid (sp.gr. 1-42) and 17.0 ml of dilute sulphuric acid (1 + l ) , evaporate the solution to fumes of sulphuric acid, fume for 5 minutes, then cool. Add 10.0 ml of 40 per cent. hydro- fluoric acid, mix, then cool the solution and transfer it into a 100-ml polythene separating * Transfer the sample solution into a 100-ml beaker before adding the tartaric acid solution. t This procedure ensures complete dissolution of these higher concentrations of niobium,138 WOOD AND JONES : SPECTROPHOTOMETRIC DETERMINATION OF [AfiahySt!, VOl. 93 funnel. Rinse the dish with 6.5 ml of water, and transfer the washings into the separating funnel. Add 2.0g of ammonium fluoride, shake gently until the reagent has dissolved, then add 25.0 ml of the equilibrated isobutyl methyl ketone (Note 6).Shake the funnel (on a mechanical shaker) for 10 minutes, then shake vigorously, by hand, for 2 minutes. Allow the organic and aqueous phases to separate, and discard the aqueous (lower) layer. To the organic extract, add 25 ml of hydrogen peroxide solution (5 volume) and shake the mixture, as before. Allow the organic and aqueous phases to separate, then run off the aqueous (lower) layer into a small platinum dish. Add 2 ml of dilute sulphuric acid (1 + l), evaporate the solution to fumes of sulphuric acid, fume for 5 minutes [if the solution darkens during evaporation, cool, add 2 ml of nitric acid (sp.gr. 1-42), and continue the evaporation], then cool. Add 2.0 ml of 50 per cent.tartaric acid solution and 5 ml of water, then transfer the solution into a 50-ml beaker with a minimum amount of water, and add 1-0 ml of fluoro- boric acid solution. Add 10.0 ml of the 5 per cent. EDTA solution, and continue as described under Colour development, Method I. NOTES- 3. Up to at least 25 per cent. of tin, 10 per cent. of aluminium, 5 per cent. of copper, vanadium, chromium, manganese or nickel, 2 per cent. of tungsten or 20 p.p.m. of tantalum do not cause a significant interference. Up to about 100 p.p.m. of tantalum can be tolerated by increasing the tartaric acid concentration (see Note 1). This extraction procedure can also be applied to the determination of niobium in molybdenum, but the procedure described in Method I11 provides a more efficient method of separating niobium from relatively pure molybdenum, and enables a larger weight of sample to be used. If alloying constituents are present, however, the solvent-extraction procedure is recommended.4. When applying this procedure to samples containing alloying constituents and 100 to 500 p.p.rn. of niobium, use a 0.1-g sample. 5. In the determination of niobium in iron (or molybdenum), cautiously add the 3 ml of nitric acid, dropwise, before adding the 40 per cent. hydrofluoric acid. 6. Partial hydrolysis of the zirconium salts occurs at this stage, but this does not affect the recovery of niobium; the insoluble material is subsequently discarded with the aqueous layer after extracting the niobium with the isobutyl methyl ketone. METHOD I11 This hydroxide-separation procedure enables the PAR method to be extended to levels of niobium from about 50 down to about 5 p.p.m. in tungsten and molybdenum (Note 7).SPECIAL REAGENTS- As described for Method I. PROCEDURE FOR NIOBIUM CONTENTS 5 TO 50 P.P.M.- Determine a reagent blank with each batch of samples, and apply the procedure to two reference solutions, each containing 2.5 ml of the standard niobium solution (1 ml of soh- tion = 10 pg of niobium). In the preparation of the reference solutions, add the standard niobium solution after the fuming stage with sulphuric acid. Transfer a 1-g sample and 10 mg of high-purity zirconium into a small platinum dish. Add 5 ml of water, 5ml of nitric acid (sp.gr. 1-42), dropwise (Note 8), and about 3 ml of 40 per cent.hydrofluoric acid, dropwise. After the sample has dissolved, cool the solution, and add 5 ml of sulphuric acid (sp.gr. 1-84). Evaporate the solution to fumes of sulphuric acid, fume for about 10 minutes, cool, add 10ml of water, then cool again. Add dilute ammonia solution (1 + l ) , dropwise, until the precipitated tungstic acid (or molybdic acid) has dissolved, and continue to add the ammonia solution until the test solution is alkaline to litmus. Allow the solution to stand for 3 hours (or overnight). Filter the precipitated hydroxides of zirconium and niobium on a No. 540 Whatman filter-paper, and wash the filter-paper and precipitate with dilute ammonia solution (1 + 99). Transfer the filter-paper and contents to a small platinum dish, dry and char, with the usual precautions, heat at 700" C for about 30 minutes, then cool.Fuse the residue with 2 g of sodium hydrogen sulphate and 1 ml of sulphuric acid (sp.gr. 1*84), then cool. Add 2.0 ml of 50 per cent. tartaric acid solution and 10 ml of water, warm gently until the melt has dissolved, then cool. Transfer the solution into a 50-ml beaker with the minimum amount of water, then add 1-0 ml of fluoroboric acid solution.March, 19681 NIOBIUM IN ZIRCONIUM, TITANIUM AND OTHER METALS 139 Add 10.0 ml of the 5 per cent. EDTA solution, and continue as described under Colour development, Method I. NOTES- 7. Based on a 0.1-g sample, no significant interference is caused by the presence of up to 1 per cent. of tin or aluminium, 0.5 per cent. of titanium, 0-1 per cent.of copper, 0.06 per cent. of iron, chromium, manganese or nickel, 0.01 per cent. of vanadium or 10 p.p.m. of tantalum. Up to about 60 p.p.m. of tantalum and 0.5 per cent. of vanadium can be tolerated by increasing the tartaric acid concentration (see Note 1). 8. In the dissolution of finely powdered molybdenum, add the nitric acid cautiously to prevent the reaction becoming too vigorous. EXAMINATION OF SAMPLES The results of the tests described below are given in Table I. TABLE I DETERMINATION OF NIOBIUM IN VARIOUS MATERIALS Hafnium Zirconium ZR20 (a) ZR30 (b) Material .. .. .. Commercially-pure titanium Titanium 318A (c) . . .. Titanium 684 (d) . . .. Titanium 679 (e) . . .. Molybdenum .. . . Tungsten .. .. .. .. .. .. .. .. .. .. .. .. .. Niobium added, p.p.m.Nil Nil 300 Nil 76 600 Nil 50 Nil 100 Nil Nil 60 200 Nil 100 Nil 200 Nil 200 Nil 50 Nil 100 Niobium determined, p.p.m. Method I Method I1 Method I11 40,46 41, 39, 42 - 30, 30 27, 27, 26 - 335, 340 - - 46,40 38, 40, 40 - 110,120 - 525, 535 - - 40,30 37,35 - 90,90 50,46 42,45 - 140, 135 - - - 1 0 , l l - - 8, 10 - - 55, 67 - - 15, 18 - - 120, 123 - - 12,16 - *(200, 220) - - - 21,24 - *(230, 220) - - - 10,12 14,14 65, 60 61, 63 60, 66 120, 110 - 100, 105 230,220 - 210,206 A f 'L - - - - *(200, 205) (a) Zirconium + 1.6 per cent. Sn, 0.12 per cent. Fe, 0.10 per cent. Cr, 0-05 per cent. Ni. (b) Zirconium + 0.5 per cent. Cu, 0-6 per cent. Mo. (c) Titanium + 6 per cent. Al, 4 per cent. V. (d) Titanium + 6 per cent: Al, 6 per cent. Zr, 1 per cent. W, 0.3 per cent. Si. (e) Titanium + 11 per cent.Sn, 2-25 per cent. Al, 6 per cent. Zr, 1 per cent. Mo, 0-3 per cent. Si. Tests made in the presence of 6 g of tartaric acid and 1 g of EDTA. The direct procedure (Method I) was applied to samples of hafnium, zirconium and zirconium alloys (ZR20 and ZR30) containing about 40 p.p.m. of niobium, and recoveries obtained were all reproducible to within about 5 p.p.m. Previous experience has shown that the acid mixture used for dissolving zirconium and hafnium does not dissolve alloying amounts of tin in ZR20, or copper in ZR30. To ensure complete dissolution of these alloys, therefore, a mixture of hydrofluoric and nitric acids was used. Samples of zirconium, hafnium, titanium, tungsten and molybdenum, to which had been added between 50 and 500 p.p.m.of niobium, were also examined by the direct procedure, and recoveries were satisfactory.140 WOOD AND JONES : SPECTROPHOTOMETRIC DETERMINATION OF [Analyst, Vol. 93 The direct procedure was also applied to samples of tungsten containing about 0-5 and 1.5 per cent. of niobium, and recoveries compared satisfactorily with those obtained with a procedure that involves the separation of niobium by hydrolytic precipitation of sodium niobatel2 and subsequent colorimetric evaluation of the niobium with hydrogen peroxide.13s1* These results are shown below. Niobium, per cent. r I Determined Determined by hydrolytic Samples Present by Method I precipitation - H,O, method Tungsten .. . . 0.6 0-53, 0.52 0.55, 0.58 1.5 1.48, 1-52 1.56, 1-59 The extraction procedure (Method 11) was applied to samples of commercially pure titanium and titanium alloys containing about 10 p.p.m.of niobium, and the recoveries were reproducible. The method was also applied to samples of titanium, to which had been added 50 and 200 p.p.m. of niobium, and good recoveries were obtained. Samples (0-5g) of zir- conium, zirconium alloys and hafnium, each taken from the same bulk material as the 0-1-g samples analysed by the direct procedure, were examined by the extraction procedure ; recoveries obtained by both methods were of the same order, but better reproducibilities were obtained by the extraction procedure. Samples of tungsten and molybdenum were examined by the hydroxide-precipitation procedure (Method 111), and reproducible recoveries were obtained.CONCLUSIONS Tests have shown that a direct procedure, based on spectrophotometric measurement of the coloured complex formed between niobium and PAR in a tartrate solution containing EDTA, can be satisfactorily applied to the determination of 50 to 500 p.p.m. of niobium in 0.1-g samples of zirconium, zirconium alloys (ZR20 and ZR30), hafnium, tungsten and molybdenum. This procedure can also be applied to 0.1-g samples of titanium and titanium alloys, provided that higher concentrations of tartrate and EDTA are used to suppress development of the titanium - PAR complex, but this modification leads to a reduction in sensitivity and limits its application to the determination of niobium in amounts above 200 p.p.m. Higher levels up to about 1 per cent. of niobium can be determined in all of these materials by dilution of the sample solution and examination of a suitable aliquot.Lower levels down to about 10 p.p.m. of niobium can be determined in titanium, zir- conium, hafnium, molybdenum and iron by a procedure that involves a preliminary extraction of the niobium from a hydrofluoric acid - sulphuric acid - ammonium fluoride solution into isobutyl methyl ketone, back-extraction of the organic phase with hydrogen peroxide solution and subsequent development of the niobium - PAR complex. Up to at least 25 per cent. of tin, 10 per cent. of aluminium, 5 per cent. of copper, chromium, vanadium, manganese or nickel, or 2 per cent. of tungsten, based on a 0.5-g sample, do not interfere. The PAR procedure can also be satisfactorily applied to the determination of down to at least 10 p.p.m.of niobium in tungsten, provided that niobium is first separated by pre- cipitation of the niobium (as hydroxide) from an ammoniacal solution, with a zirconium hydroxide carrier. The same procedure is equally applicable to the determination of niobium in molybdenum; separation of these two metals in this way is more efficient than by the extraction procedure, and it enables a larger weight of sample to be used. Solvent extraction is advantageous, however, when certain alloying constituents are present. Because of the close chemical similarity of niobium and tantalum, these metals cannot be individually identified, either by incorporating an isobutyl methyl ketone extraction stage, or following a hydroxide precipitation.Like niobium, tantalum also forms a coloured complex with PAR under the specified conditions, but the sensitivity of the tantalum - PAR complex is only about one eighth of that of the corresponding niobium complex; hence, the effect of tantalum in the determination of 50 to 200 p.p.m. of niobium, based on a 0.1-g sample, is only likely to be significant when tantalum is present in excess of 50 p.p.m. The effect of tantalum can be reduced by increasing the concentration of tartaric acid; a similar amount of tartaric acid must also be present in the reference solution, because an increase in tartrate concentration also causes a decrease in the sensitivity of the reagent to niobium.March, 19681 NIOBIUM I N ZIRCONIUM, TITANIUM AND OTHER METALS 141 Because variations in the concentrations of the reagents and the quality of the PAR have a significant effect on the absorbance of the niobium - PAR complex, it is recommended that the final absorbance of the sample solution be related to absorbance obtained on standard solutions of niobium that have been simultaneously examined by the entire procedure. With the direct procedure, a batch of about twelve determinations can be completed in a normal working day; when a preliminary separation involving either extraction or pre- cipitation of the niobium is necessary, the number of determinations is reduced to about eight. The authors appreciate the interest shown by Professor R. Belcher, and thank Mr. W. T. Elwell 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. for helpful-advice and assistance in t h e preparation of this paper. REFERENCES Klinger, P., and Koch, W., Arch. EisenhiittWes., 1939, 13, 127. Platonov, M. S., Krivoshlykov, N. F., and Maraka’yev, A. A., Zh. Obshch. Khim., 1936, 6, 1815. Kassner, J. L., Garcia-Porrata, A., and Grove, E. L., Analyt. Chem., 1955, 27, 492. Lauw-Zecha, A. B. H., Lord, S. S., and Hume, D. N., Ibid., 1952, 24, 1169. Cheng, K. L., and Goydish, B. L., Talanta, 1962, 9, 987. Ackerman, G., and Koch, S., Ibid., 1962, 9, 1015. Belcher, R., Ramakrishna, T. V., and West, T. S., Ibid., 1962, 9, 943. Elinson, S. V., Pobedina, L. I., and Rezova, A. T., Zav. Lab., 1966, 32, 1314. Alimarin, I. P., and Hsi-i Han, Zh. Analit. Khim., 1963, 18, 182. Belcher, R., Ramakirshna, T. V., and West, T. S., Ckem. & Ind., 1963, 531. Milner, G. W. C., Barnett, G. A., and Smales, A. A., Analyst, 1955, 80, 380. Schoeller, W. R., and John, C., Ibid., 1927, 52, 513. Wood, D. F., and Adams, M. R., Analytica Chim. Acta, 1964, 31, 153. Elwell, W. T., and Wood, D. F., “Analysis of the New Metals,” Pergamon Press, Oxford, 1966, Received August 30th, 1967 p. 170.