Determination of traces of vanadium in molybdenum metal and

(4) Fell, A. F.; Clark, B. J.; Scott, . P. J. Chromatogr. 1984, 316, 423. (5) Webb, P. A.; Ball, D.; Thornton, T. J. Chromatogr. Scl. 1983, 21,. 447. ...
0 downloads 0 Views 650KB Size
Anal. Chem. 1987, 59, 1907-1911

collection of chromatographicand spectral data. We gratefully acknowledge many helpful discussions with R. King, G. Verghese, R. Kim, J. Hallowell, S. Rhode, and D. Mauro. Our success with the VOL criteria is due in part to encouragement and practical suggestions received from G. Verghese, for which we are particularly appreciative.

LITERATURE CITED Osten, D. W.; Kowalskl, 8. R. Anal. Chem. 1984, 56. 991. Vandeglnste, B.;Essers, R.; Bosman, T.; Reijnen, J.; Kateman, G. Anal. Chem. 1985, 5 7 , 971. Carter, G. T.; Schlesswohl, R. E.; Burke, H.; Yang, R. J . Pharm. Sci. 1982, 71, 317. Fell, A. F.; Clark, B. J.; Scott, H. P. J . Chromatogr. 1984, 376, 423. Webb, P. A.; Ball, D.; Thornton, T. J . Chromatogr. Sci. 1993, 21, 447. Cheng, H.; Gadde, R. R. J . Chromatogr. Sci. 1985, 23, 227. Ll, J.; Hllller, E.; Cotter, R. Am. Lab. (FalrfleM. Conn.) 1985, 17, 93. LI, J.; Hillier, E.; Cotter, R. J . Chromatogr. Sci. 1985,,23, 446. Drouen, A. C. J. H.; Bllliet, H. A. H.; DeGalan, L. Anal. Chem. 1984, 56, 971. Drouen, A. C. J. H.; Bllllet, H. A. H.; DeGalan, L. Anal. Chem. 1985, 57, 962. Fell, A. F.; Clark, B. J.; Scott, H. P. J . Pharm. Biomed. Anal. 1984, 1 , 557. Fell, A. F.; Scott, H. P.;Gill, R.; Moffat, A. C. J . Chromatogr. 1983, 273, 3. Baker, J. K.; Skekon, R. E.; Ma. C. J . Chromatogr. 1979, 168, 417. Krstulovlc, A. M.; Brown, P. R.; Rosie, D. M. Anal. Chem. 1977, 49, 2237. Hartwick, R. A.; Assenza, S. P.; Brown, P. R. J . Chromatogr. 1979, 186, 847. Yost, R.; Stoveken, J.; MacLean, W. J . Chromatogr. 1977, 134, 73.

1907

(17) Cohen, S. A. Pittsburgh Conference and Exposition, New Orleans, LA, 1 8 8 5 paper 320. (18) Warren, F. V., Jr.; Bldilngmeyer. B. A.; Delaney, M. F. Anal. Chem., preceding paper In this Issue. (19) Warren, F. V., Jr. Ph.D. Thesls, Boston University, 1986. (20) Malinowski, E. R. Anal. Chim. Acta 1982, 734, 129. (21) Mallnowski. E. R.; Howery, D. 0.Factor Analysis in Chemistry;Wlley: New York, 1980; Chapter 4. (22) Mallnowskl, E. R. Anal. Chem. 1977, 4 9 , 612. (23) Wold, S. Technometrlcs 1978, 2 0 , 397. (24) Strang, 0. Linear Algebra and Its Applications: Academic: New York, 1980; Chapter 3. (25) Junker, A.; Bergmann, G. Z . Anal. Chem. 1974, 272, 267. (26) Massart, D. L.; Dljkstra, A.; Kaufman, L. Evaluation and Optlmlzation of Laboratory Methods and Analytical Procedures ; Elsevler: Amsterdam, 1978; Chapter 17. (27) ZuDan, J. clustering of Large Data Sets; Research Studies Press: Chichester, 1982. Bevlngton, P. R. Data Reduction and Error Analysis for the Physical Sciences; McGraw-HIII: New York, 1969. Kateman, 0.; Pijpers. F. W. Qual& Control in Analytical Chemistry; Wiley: New York, 1981; Chapter 4. Warren, F. V.; Phoebe, C. H.; Bldllngmeyer, 8. A. Ninth International Symposium on Column Liquid Chromatography, 1985, Edinburgh, U. K., poster P03.11. Drouen, A. C. J. H.; Bllliet, H. A. H.; Schoenmakers, P. J.; dealan, L. Chromatographia 1982, 16, 48. Mallnowskl. E. R.; Cox, R. A.; Haldna, U. L. Anal. Chem. 1984, 56, 778. (33) Mallnowski, E. R., Stevens Institute of Technology, Hoboken, NJ, personal communlcation, 1985.

RECEIVED for review April 29, 1986.

Resubmitted December

2, 1986. Accepted March 17, 1987.

Determination of Traces of Vanadium in Molybdenum Metal and Compounds by Ion Exchange Chromatography-Spectrophotometry F. W. E. Strelow National Chemical Research Laboratory, CSIR,

P.O.Box 395, Pretoria 0001, Republic of South Africa

Down to sub-parts-per-million concentrations of vanadium have been determined In up to 5 g of molybdenum metal or 10 g of its compounds. The vanadium has been separated by cation exchange chromatography in citric acld on 3-g AG50W-XS resin columns, followed by spectrophotometry of the 4-(2-pyrldylaro)resorclnol (PAR) complex in the presence of CDTA (cyciohexanedlamlnetetraacetlc acid) to complex possible interfering trace elements. Separations are quite satisfactory, leavlng only between 5 and 30 pg of molybdenum in the vanadium fractlon when 10 g of H,MoO, was present originally. Recoveries of vanadlum from synthetic mlxtures are quantitative, and 5-pg amounts of vanadium in 10 g of H,MoO, have been determined with a standard deviatlon of 0.05 pg. Distribution coefficients of vanadium( I V ) and iron( I I I ) in cltric and citric-nitric acid mixtures as well as in mixtures containing phosphoric acid are presented together with elution curves showing that even large amounts of vanadium can be retained effectively on a 3-g column. Resuns for the analysis ot synthetk mixtures and some actual samples are included, and the Influence of the concentrations of nitric, citric, and phosphoric acld on the separation is dlscussed.

I t has been shown recently that traces of titanium down to sub-parts-per-million levels in molybdenum metal and compounds can be determined by a combination of cation 0003-2700/S7/0359-1907$01.50/0

exchange chromatography and spectrophotometry of the titanium-Tiron complex ( I ) . A large number of other trace elements such as iron, copper, manganese, lead, etc. can be determined by combining cation exchange chromatography with flame atomic absorption (2) or X-ray fluorescence (3). When the separation of traces of vanadium from large amounts of molybdenum is required, the above ion exchange procedures (1-3),which use a cation exchange resin, are not suitable, because in the peroxide-containing solutions obtained after the dissolution step, vanadium is present as an anionic vanadium(V) peroxide complex, which passes through the column together with molybdenum. Klement (4)has described a cation exchange procedure that uses citric acid instead of hydrogen peroxide to prevent precipitation of about 100-mg amounts of molybdenum. Vanadium is reduced to the four-valent state with sulfur dioxide and adsorbed as a cation. Only relatively large amounts of vanadium(IV), between 74 and 297 mg, were separated from binary mixtures with molybdenum and determined by titration with permanganate. No attempt to separate and determine microgram amounts of vanadium was made. Flame atomic absorption is not very sensitive for the determination of vanadium and for this reason not suitable for the determination of amounts that are less than 20 bg in 10 mL solution, while furnace atomic absorption, which is more sensitive, is considerably less accurate and more time-consuming. 4-(2-Pyridylazo)recorcinol(PAR) is one of the most sensitive reagents for the spectrophotometric determination 0 1987 American Chemical Society

1908

ANALYTICAL CHEMISTRY, VOL. 59, NO. 15, AUGUST 1, 1987

of vanadium (5) and has been used successfully for the accurate determination of vanadium in international reference materials (rocks) down t o a few parts per million (6). The colored complex has a molecular extinction coefficient of 3.6 x lo4, and 1 ppm of vanadium in solution produces an absorbance of 0.630 at 545 nm. This paper combines a revised and adapted cation exchange procedure in citric acid with spectrophotometry using PAR for the accurate determination of vanadium down to subpart-per-million levels in molybdenum metal and some compounds. For larger amounts (500 pg or more) flame atomic absorption is faster and as accurate. Because the concentration of citrate and of hydrogen ions can have a pronounced influence on the distribution coefficients and column behavior of vanadium(1V) such information is presented. Furthermore, similar information for iron(III), a typical multivalent cation, is included to show that such elements are also retained very strongly. In addition, systematic information about the influence of phosphoric acid on the ion exchange behavior of vanadium(1V) is included, because phosphoric acid is employed in some dissolution procedures (3). The developed method is applied to the determination of vanadium in binary synthetic mixtures with molybdenum, in molybdenum metal, and in two compounds.

EXPERIMENTAL SECTION Reagents. Reagents were of AR grade purity unless stated otherwise, and distilled water was furth r purified by passing through an Elgastat deionizer. A standara solution of vanadium was prepared by dissolving 12.65 g of Merck vanadium(IV) oxide sulfate “pure” (VOS04.5H20)in deionized water and diluting to 500 mL (0.100 M vanadium). The exact vanadium concentration was determined by titration of suitable aliquots with potassium permanganate. A solution containing exactly 500 ppm of vanadium in 0.5 M nitric acid was prepared from the above by appropriate dilution and addition of a calculated amount of nitric acid. Solutions containinglower concentrationsof vanadium were prepared by further dii Jtion when required. Vanadium-free solutions of molybdenum for quantitative separations of synthetic mixtures were obtained by dissolving 10-g amounts of molybdic acid (H2Mo04)in 30 g of citric acid plus 200 mL of deionized water by heating on the water bath. After removal of remaining minute traces of undissolved material (less than 1 mg) by filtration, some deionized water containing SOz was added. The cold solutions, diluted to about 300 mL, were passed through columns containing 15 mL of AG50W-X8 cation exchange resin (70-mm length and 16.5-mm diameter). This removed not only the last traces of vanadium but also most other cationic elements except lithium. The resin was the AG50W-X8 sulfonated polystyrene cation exchanger of 100-200 mesh particle size marketed by BIO-RAD Laboratories of Richmond, CA. Columns and Apparatus. Borosilicateglass tubes of 15-mm i.d. and 300-mm length, fitted with a no. 1 porosity glass sinter plate and a buret tap at the bottom and a B14 female ground glass joint at the top, were used as columns (3 g of resin). In some cases (5and 10-gresin columns) glam tubes of 16.5-mmid. and W m m length fitted with a B19 female ground glass joint at the top were used. The columns were filled with a slurry of AG50W-XS resin of 100-200mesh particle size until the settled resin reached a 9.0-mL mark (3 g of resin). The resulting resin columns were 52 mm long (in H20). In some cases (elution curves in presence of phosphoric acid) wider columns (16.5-mm i.d.) were used and filled to a 15or 30-mL mark (5 or 10 g of resin). In these cases the resulting resin columns were 70 and 140 mm long, respectively. The resin was purified by passing through about 100 mL of 5 M hydrochloric acid (200 mL for the larger columns), followed by 50 mL of deionized water (100 mL for the larger columns). Spectrophotometric measurements (vanadium) were carried out with a Zeiss PMQ I1 UV-visible spectrophotometer with 10-mm cells. A Varian-Techtron AA-5 instrument was used for atomic absorption measurements (molybdenum, vanadium, and iron).

Table I. Distribution Coefficients in Citric Acid-HNO, molarity of HNOB

molarity of citric acid

nil nil nil nil

0.05 0.10 0.25 0.50

nil 0.10 0.10 0.10 0.10

0.10 0.20 0.20 0.20 0.20 0.20 0.50 0.50 0.50 0.50 0.50

D[vanadium(IV)] D[iron(III)] 74 000 45 000

3500 3500 1510

21 000 17 000 4 300 22 000

850 560

1.00

0.05

630 600 520 465 330 239 233 223 198 177 56 56 56 55 53

0.10

0.25 0.50 1.00 0.05 0.10

0.25 0.50 1.00

0.05 0.10

0.25 0.50 1.00

18000 10000 5 700 2 770 7 000

6 200 4 160 2 570 1290 800 760 670 550

426

Table 11. Variation of Distribution Coefficients of V(1V) with H3P0, Concentrations” molarity

molarity

of

of H3P04

H3P04 D[vanadium(IV)]

nil 0.02 0.05 0.10

485 550

454 332

0.20 0.50 1.00

D[vanadium(IV)] 204 81

34.0

“The molarity of HN03 = 0.10 and the molarity of citric acid = 0.50 for all equilibria.

Distribution Coefficients in Citric Acid-HNOB. Portions (2.500 g) of AG50W-X8 resin (dried at 110 “C) were equilibrated in a mechanical shaker for 24 h at 20 OC with 250 mL of a solution containing either 1.00 mmol of vanadium(1V) or 0.67 mmol of iron(II1) and citric and nitric acid in concentrations ranging from 0.05 to 1.00 M and from nil to 0.50 M, respectively. The actual concentrations of both acids are shown in Table I. After equilibration the resin was separated from the aqueous phase by filtration on a 200-mm-long and 23-mm-wide (id.) glass column with a sinterplate of no. 1 porosity at the bottom. The resin was transferred into the columns and washed with deionized water, and the aqueous phase was then kept separately. The adsorbed part of the element was eluted with about 3 M hydrochloric acid and also kept apart. The amounts of the elements in both phases were determinedby flame atomic absorptionafter suitable dilution and the distribution coefficients calculated from the analytical resulh. The coefficients are presented in Table I. Effect of &PO, on Ristribution Coefficients of V(1V) in Citric Acid-HN03. Table I1 shows distribution coefficients of vanadium(1V) in 0.50 M citric acid plus 0.10 M nitric acid containing various amounts of phosphoric acid ranging from nil to 1.00 M. The amounts of vanadium again were 1.00 mmol per equilibrium and the experimental conditions and procedure were similar to those described above. Elution Curve. A solution containing 3.33 g of molybdenum (5 g M d J , 10 mg of vanadium(IV),and 20 g of citric acid in about 200 mL was passed through a column containing 3 g (9 mL) of AG50W-X8 resin as described above. The column had been equilibrated by passing through about 20 mL of 0.1 M citric acid containing 0.01 M nitric acid. The solution was washed onto the resin with a few small portions of the same reagent, which was then also used to elute molybdenum at a flow rate of 3.0 f 0.5 mL/min. Though elution of molybdenum was complete after 40 mL (