Anion Exchange Separation .of Zirconium, Titanium, Niobium, Tantalum, Tungsten, and Molybdenum W. R BANDI, E. G. BUYOK, 1. L. LEWIS, and L. M. MELNICK Applied Research Laboratory, United States Steel Corp., Monroeville, Pa.
b Many modern alloys contain zirconium, titanium, niobium, tantalum, tungsten, and molybdenum. The determination of these elements in mixtures has long been regarded as one of the most difficult problems in inorganic analysis, because the necessary separations are made with difficulty. The separation problems have been met in the last decade by application of ion exchange chromatography techniques with hydrochloric-hydrofluoric acid media. However, these separation schemes have not been wholly acceptable because of the inherent difficulties associated with working with large volumes of concentrated hydrofluoric acid. The anion exchange behavior of these six metals in solutions that contained two or more of the reagents hydrochloric acid, oxalic acid, citric acid, ammonium chloride, ammonium citrate, and hydrogen peroxide has been studied, and separation conditions have been established. Information on column size and volumes required for the selective elutions, as well as analytical results on synthetic and standard samples, is presented.
B
of the increasing use of complex alloy steels and hightemperature alloys for supersonic aircraft and guided missiles, there is a need for better methods for the analysis of alloys containing titanium, vanadium, zirconium, niobium, tantalum, tungsten, and molybdenum. The problems that are encountered in the gravimetric separation of these elements are well known (7, IS). In recent years, other techniques have been applied to the separation of these elements-via., liquidliquid extraction (9, 12, 16), distillation (1, 8 ) , columnar cellulose chromatography (9, 1 4 , and ion exchange chromatography (4,6, 11). Of these newer separation techniques] ion exchange chromatography has been the most successful. Kraus and coworkers (10, 11), Hague and coworkers (6, 6), and Wilkins (18) have studied the separation of certain of the above elements by anion exchange chromatography with eluents containing hydrofluoric and hydrochloric acids. The use of high concentrations of hydroECAUSE
fluoric acid, however, has certain d i 5 advantages: the hazard of hydrofluoric acid burns, the need for using laboratory equipment of plastic or platinum, and the additional work needed to separate fluoride from some of the metals before the final determinations. Wacker and Baldwin (17) have studied the ion exchange separation of zirconium and titanium in solutions of oxalic and hydrochloric acids, and Speecke and coworkers (3, 4, 15) have studied the separation of niobium and tantalum in solutions of the same acids. These studies indicated that useful analytical separations for all six elements-zirconium, titanium] niobium, tantalum, tungsten, and molybdenumcould possibly be made in oxalic acidhydrochloric acid media if more were known of the variables that might affect these separations. The success of the anion exchange separation of these metals would depend on forming anionic species of diverse charge, size, and stability. By maximizing the differences existing between the different metals, the separation process would be simplified. Inasmuch as oxalic-hydrochloric acid has been used successfully, it seemed that other complexing agents--e.g., peroxide, citrate, and pyrogallol- could lead to further improvement. hZoreover, the use of these complexing agents in the eluent would have the advantage that some of the metals could be determined directly after the separation. For example, titanium, tantalum, and niobium react with pyrogallol in the presence of oxalate to form colored substances. MATERIALS AND EXPERIMENTAL W O R K
Materials. IONEXCHANGE COLA slurry of Dowex 1-X8 resin, 200- to 400-mesh, was made by adding distilled water to the resin. Any fine particles that did not settle within 10 minutes were decanted and discarded. The slurry was transferred to a borosilicate glass column 15 cm. long and 1.8 cm. in inside diameter until a resin bed height of 10 cm. was obtained. A perforated bottom from a porcelain Gooch crucible was added to the top of the resin to prevent disturbance of the resin upon introduction of solutions. A 500-ml. separatory UMN.
funnel was attached to the top of the column with a rubber stopper and glass and rubber tubing. This funnel served as a reservoir for the various eluting agents. New resin was cleaned by eluting alternately with 9M hydrochloric acid and water until the eluent was colorless. The column was preconditioned by elution with 50 ml. of 1.5-44 hydrochloric acid-0.5M oxalic acid solution. REAQENTS.Titanium, zirconium, niobium, tantalum, tungsten, and molybdenum are usually isolated from the matrix material as insoluble acids or insoluble organic chelates. After these compounds are ignited t o their oxides, the oxides are fused in bisulfate to obtain the elements in a soluble form. In this work the elements were finally dissolved in oxalic acid. Because the ion exchange separations would be affected by the sulfate and oxalate present, the experimental separations were made on solutions treated similarly. All solutions were prepared to contain 1.00 mg. of metal per ml. of solution. TITANIUMSOLUTION.Titanium dioxide (0.3349 gram of sample 154a, National Bureau of Standards) was fused with 8 grams of potassium bisulfate, and the resulting melt was dissolved in 100 ml. of 6.5H oxalic acid and diluted to 200 ml. with water. ZIRCONIUM SOLUTION. &agent grade anhydrous zirconium dioxide (0.2702 gram) was fused with 8 grams of potassium bisulfate. The resulting melt was dissolved in 50 ml. of 0.5M oxalic acid and diluted to 200 ml. with distilled water. NIOBIUMSOLUTION.Exactly 0.2000 gram of reagent grade niobium metal (99.5% mininum) was ignited to the oxide and treated in the same manner as the titanium dioxide. TANTALUM SOLUTION. Tantalum foil [0.2000 gram of reagent grade tantalum (99.9% minimum)] was treated in the same manner as the niobium. MOLYBDENUM SOLUTION.For the standard molybdenum solution, 0.3016 gram of reagent grade molybdenum trioxide (99.5% minimum) was treated in the same manner as the titanium dioxide. TUNGSTENSOLUTION. Exactly 0.2522 gram of tungsten trioxide (obtained by dehydrating NazWOA with hydrochloric and nitric acids and igniting the precipitate) was treated in the same manner as the titanium dioxide. VOL. 33, NO. 9, AUGUST 1961
1275
VANADI~YM SOLUTION. For this solution, 0.3568 gram of reagent grade vanadium pentoxide was treated in the same manner aa the titanium dioxide. Experimental Procedure. ANION EXCHANGE SEPARATION OF ELEMENTS.
Elution data were obtained
on columns' that were pretreated with the eluent solution being tested, and to which was then added 10 mg. of one of the six elements. Only one of the metals was added to each column, and data for all six metals were obtained with each eluent solution. The eluates were collected in 30-ml. fractions and analyzed semiquantitatively for the metal ion by a standard colorimetric method. In the separation studies made, the concentrations of the components of the eluents were varied over rather broad .ranges. The following anion exchange procedure was developed as a result of these studies: Transfer a bisulfate-0.5M oxalate solution (30 to 100 ml.) containing the six elements to a column prepared as described. To the reservoir, add 200 ml. of a solution that is 1.5M in hydrochloric acid, 0.5M in oxalic acid, and 0.007M in hydrogen peroxide. Collect the eluate in a flask or beaker at a flow rate of 2 to 3 ml. per minute. (The flow rate can be regulated with a stop cock or screw clamp.) Reserve this eluate for the determination of titanium and zirconium. Continue the elution with an additional 550 ml. of the same eluent solution and reserve this eluate for the determination of niobium. Transfer 300 ml. of an eluent solution that is 3M in hydrochloric acid and 0.5M in oxalic acid to the reservoir and collect the eluate fraction for the determination of tantalum. Add 600 ml. of an eluent that is 4M in hydrochloric acid and 0.1M in citric acid to the reservoir and collect the eluate for the determination of tungsten. Finally, elute with 200 ml. of a solution 1.9M in ammonium chloride and 0.44.44 in ammonium citrate; reserve this eluate fraction for the determination of molybdenum.
0 Eluent Solution
'.5 M Hydrochloric Acid 0.5 M oxalic Add 0.007 M HydrO9m Peroaidr
\
\
100
Figure 1. H202
200
3w
400 500 ELUATE VOLUME, ml
600
800
700
900
Elution behavior of the elements with 1.5M HCI and with and without
A Eluenl Solution
2.0 M HydroChloric Acid 0.5 M Oxalic Acid
I 0. Eluenl solution
2.0 M Hydrochloric Acid 0.5 Y OIaIK Acid 0.007 M Hydroqtn Peroxide
'\"
I
100
Figure 2. H202
200
300
1276
ANALYTICAL CHEMISTRY
600
700
000
900
Elution behavior of the elements with 2.OM HCI and with and without
A. Eluent Solution
3.0 M Hydrochloric Acid 0.5 M Oxolic Acid
.....
ANALYSISOF ELUENT FRACTIONS. Because many established methods exist for the determination of these metals after the necessary separations have been accomplished, the methods used for the evaluation of the separations are summarized only briefly. TITANIUMAND ZIRCONIUM. After the oxalic acid was removed by volatilization, the metals were precipitated with cupferron. Titanium was determined spectrophotometrically with hydrogen peroxide, and zirconium was determined either gravimetrically with benzenearsonic acid or photometrically with chloroanilic acid. NIOBIUMAND TANTALUM. The respective eluates were adjusted to the proper pH with ammonia, and niobium and tantalum were precipitated with tannic acid. Any silicon was volatilized
400 500 ELUATE VOLUME, ml
0. Elvrnl M u t i o n 3.0 M Hydrochloric Acid 0.5 M Oaalic Acid 0.007 M Hydrogen Peroxide
\,Ta
'. IO0
Figure 3. H202
200
300
400 500 ELUATE VOLUME. ml
600
700
800
900
Elution behavior of the elements with 3.OM HCI and with and without
with hydrofluoric and sulfuric acids and the residue was reweighed except when the concentration of tantalum was low. In these instances tantalum was determined photometrically with pyrogallol.
TUNQSTEN. Citrate was destroyed by fuming with nitric and sulfuric acids, and tungsten was determined photometrically with hydroquinone or gravimetrically with cinchonine.
0 466
0 400
300 ml
600 ml
3 hi Hydrochloric Acid 0.5 M Oxalic Acid
4 M Hydrochloric Ac!d 0.I M Citric Acid
.i ! 19 M Ammonium 200 mi
*
j
Chloride
0 44 M Ammonium
0.333
5 D
.E 0.266
z
P
5
+ z
E
0.200
z u 0 0
0.I33
0.066
I .....*.' I
0,
1200
1400
1 : 1600
ELUATE VOLUME, ml
Figure 4.
Anion exchange s e p a r a t i o n of 20 mg. each of Ti, Zr,
MOLYBDENUM. After the citrate ion was destroyed, molybdenum was determined photometrically with thiocyanate or gravimetrically with abenzoinoxime. RESULTS AND DISCUSSION
Effect of Hydrochloric Acid Concentration. The hydrochloric acid
concentration had a marked effect on the elution characteristics of the elements. None of the elements was eluted with 250 ml. of 0.5M oxalic acid, or with 250 ml. of 0.5M hydrochloric acid-0.5M oxalic acid. However, when the concentration of hydrochloric acid was increased to 1.5M, zirconium was eluted completely within the first 200 ml. of eluent. This is shown in Figure 1,A. On continued elution, niobium was removed next with the same solution. Figure l , A , also shows that with this eluent, titanium and tantalum overlapped the elution of niobium. No elution of tungsten or molybdenum was observed with 660 ml. of the 1.5M hydrochloric acid-0.M oxalic acid eluent solution. When the hydrochloric acid concentration was increased from 1.5 to 2.0MJ zirconium, niobium, titanium, and tantalum were eluted rapidly (Figure 2,A); however, a less satisfactory separation of the elements occurred. Tungsten and molybdenum still were not eluted. B further increase in the hydrochloric acid concentration to 3M resulted in incomplete resolution, with some of
the niobium, titanium, and tantalum being eluted simultaneously (Figure 3, A). Tungsten was eluted in this experiment but molybdenum waa not. A 4M hydrochloric acid-0.5M oxalic acid solution showed elution behavior for all elements similar to that for the 3M hydrochloric acid-0.5M oxalic acid solution. It, therefore, appeared that certain of the elements could be separated by using large volumes of eluent that contained low concentrations of hydrochloric acid, and by using longer ion exchange columns; however, such changes might have increased separation time to the extent that the procedure might not have been attractive. EFFECTOF ADDITIONOF 0.007M HYDROGEN PEROXIDE. To facilitate the ion exchange separation of the six ele,ments in hydrochloric-oxalic acid medium, the addition of reagents that would tend to change the ionized forms of the metal ions on the column wm attempted. Initial experiments with dilute hydrogen peroxide showed that it had a pronounced effect on the elution behavior of the elements. In addition, hydrogen peroxide causes yellow bands of titanium, niobium, and molybdenum to form on the column, which provide a visible means of following the progress of separations. As shown by comparing Figure 1,A, with 1,B, the elution of titanium is markedly accelerated by adding 0.007M hydrogen peroxide to a 1.5M hydro-
Nb, Tal W,
and Mo
chloric acid-0.5M oxalic acid eluting solution. Also, hydrogen peroxide appeared to cause tantalum to be held more tightly on the resin. Niobium elution was not affected greatly by the addition of peroxide. Figure 2,A and B shows the difference in elution observed on adding hydrogen peroxide to a 2M hydrochloric acid-0.5M oxalic acid eluting solution. Again, the acceleration of titanium elution is noted, and the retardation of tantalum elution is pronounced. Figure 3, A and B, shows a difference in the elution behavior of titanium and tantalum similar to that shown in Figure 2,A and B. In addition, Figure 3,B shows that the elution of tungsten is retarded by hydrogen peroxide. Separation of Metal Ions. The results of an attempt to separate zirconium, titanium, and niobium by varying the hydrochloric acid concentration and using hydrogen peroxide in the first 750 ml. of eluting solution are shown in Figure 4. Titanium and Zirconium can be eluted in the initial fraction, and niobium in a second fraction. If i t is necessary to separate titanium and zirconium, this can be accomplished by destroying their peroxy complexes, transferring the titanium-zirconium fraction to another column, and eluting the zirconium with 1.5M hydrochloric acid-0.5M oxalic acid solution containing no peroxide. Subsequently, the titanium can be eluted with 100 ml. of 3M hydroVOL. 33, NO. 9, AUGUST 1961
* 1277
Table 1. Recovery of Elemenis after Separation by Ion Exchange
20 Mg.
Element Zirconium Titanium Niobium Tantalum Tungsten Molybdenum
Recovered % 100 99.5 104.5 103 100 97
20.0 19.9 20.9 20.6 20.0 19.4
be expected as a major or minor constituent in any of the materials that may contain a combination of the other elements studied, and because an effort to separate vanadium from the other elements before applying ion exchange would be cumbersome, i t was necessary to establish the elution behavior of vanadium in this separation scheme. Vanadium was eluted with the zirconium-titanium fraction and the presence of vanadium was indicated by a blue band.
LITERATURE CITED
( 1 ) Atkinson,
R. H., Steigman, J., Hiskey, C. F., ANAL.CHEM.24, 447811952). >----,
(2) Burstall, F. H., Williams, A. F., Analyst 77, 983 (1952). (3) Gilles, J., Eeckhout, J., Cornand, P., Sueecke, A.. Mededel. Koninkl. Vlaam. Acad. Wetenschap. Belg., K1. Wetenschap 15, No. 13,3-17 (1953). (4) Gilles, J., Hoste, J., Cornand, P., Speecke, A., Mededel. Vlaam. Chem. Ver. 15, 63-5 (1953). (5) Hague, J. L., Brown, E. C., Bright, H. A., J. Research Natl. Bur. S t u d ai-& 53,261-2 (1954),
Table II. Analysis of National Bureau of Standards Samples
Standard 123ba Found with ion Element exchange sepn. Nb 0.72, 0.73, 0.74 Ta 0.20. 0.18. 0.19 W 0.18;0.18; 0 . 1 8 Mo 0.17, 0.17, 0.17 Ti 0.008, 0.008 V ... Sample weight, 3 g. b Sample weight, 0.4 g.
Av. 0.73 0.19 0.18 0.17 0.008
...
chloric acid-0.5M oxalic acid-0.007M hydrogen peroxide. Tantalum can be eluted in a third fraction by increasing the hydrochloric acid concentration and by eliminating the hydrogen peroxide in the eluent. Figure 4 shows a double maximum in the elution of the tantalum. This and other spectrophotometric evidence compiled at this laboratory lead one to believe that there exist two oxalate complexes of tantalum in acid solution. (The single elution peak for tantalum in Figure 3 4 , is attributed to the fact that the data were determined semiquantitatively.) All attempts to elute tungsten by increasing thc hydrochloric acid concentration and eliminating the hydrogen peroxide led to tailing and incomplete removal of the tungsten from the column. Evidently, the hydrogen peroxide formed a stable complex with tungsten, which was tightly absorbed by the resin. Tungsten was finally eluted with 4M hydrochloric acid-0.1M citric acid solution (see Figure 4). The elution of molybdenum was attempted with solutions of the following compounds: sodium oxalate, sodium thiocyanate, sodium hydroxide, sodium acetate, ammonium tartrate, ammonium oxalate, ammonium chloride, and ammonium citrate. Figure 4 shows that molybdenum was successfully desorbed by eluting with a 1.9M ammonium chloride-0.44M ammonium citrate solution. Behavior of Vanadium in Separation Scheme. Because vanadium could 1278
0
ANALYTICAL CHEMISTRY
Standard 168b Tentative Found with ion value exchange sepn. 0.75 0.20 0.18 0.17 0.006
...
2.95, 2.95 0.92. 0.98 3.95; 3.95 3.90, 3.88 0.05, 0.05 0.03’
Av. 2.95 0.93 3.95 3.89 0.05 0.03
Evaluation of Separation. Table
I shows the per cent recovery of the elements after separation of 20-mg. amounts of each. These values were obtained by summing the results obtained on successive 30-ml. fractions of eluate, the same data being used to plot Figure 4. Therefore, the deviations from 100% are most probably due to the cumulative errors of a number of analyses rather than to the errors of the separation. Mixtures of 200 mg. of these elements were separated when the peroxide concentration was increased; however, the optimum peroxide concentrations were not established. The completeness of the separation of the elements is also demonstrated by the results obtained on the analysis of National Bureau of Standards (NBS) steel samples. The tentative standard values and the results obtained by this method are listed in Table 11. The agreement is good. Initially low values for tungsten and molybdenum were obtained for NBS samples 167 and 168. However, these low results were found to be due to the incomplete isolation of the tungsten and molybdenum oxides from the standard samples. It was determined that the oxides of all the metals could be isolated by using cinchonine, cupferron, and a-benzoinoxime in three separate precipitations. The precipitates were then combined, ignited, fused, and leached in oxalic acid prior to the ion exchange separation.
Standard 167* Tentative value 2.95 0.. 9_. 5 3.95 3.95 0.06 0.03
Found with ion exchange sepn.
Tentative value
3.15 0.07 4.52 3.87
4.50
Not determined
0.01
...
3.15
n.nR -.--
3.90
...
(6) Hague, J. L., Machlan, L. A., ZMd., 63,Il-19(1959). (7) Hildebrand, W. H., Lundell, G. E. F., Bright, H. A., Hoffman, J. I plied Inorganic Analysis,” 21;)d p. 588, Wiley, New York, 1953. (8) Hiskey, C. F., Newman, L., Atkinson, R. H., ANAL.CHEM.24,1988-91 (1952). (9) Kassner, J. L., Garcia-Porrata, A,, Grove, E. L., ZMd., 2 7 , 4 9 2 4 (1955). (10) Kraus, K. A., Moore, C. E., J . Am. Chem. Soc. 71,3855 (1949). (11) Kraus, K. A., Nelson, F., “Anion Exchange Studies of the Fission Producta,” Vol. VII, Proceedings of International Conference on Peaceful Uses of Atomic Energy, Geneva, 1955, United Nations, New York, N. Y., 1956. ZjMilner, G. W. C., Barnett, C. A., Smalea, A. A., Analyst 80, 380-90 (1955). 3) Schoeller, W. R., “Analytical Chemistry of Tantalum and Niobium,” p. 1, Cha man and Hall, London, 1937. 4) Schoezr, W. R., Powell, A. R., “Analysis of Miner,r4ls and Ores of the Rarer Elements, 3rd ed., p. 210, Hafner Publishing Co., New York, 1955. 5) Speecke, A., Hoste, J., Talanta 2, 332-40 (1959). 6) Stevenson, P. C., Hicks, H. G., ANAL.CHEM.25, 1517 (1953). 7) Wacker, R. E., Baldwin, W. H.,
“2
“Se aration of Zirconium and Niobium hayate Complexes by Anion Exchange,” Technical Information Division, U. S . At. Energy Comm., Oak Ridge National Laboratory, Oak Ridge, Tenn., paper 637, June 21, 1950. (18) Wilkins, D. H., Talanta 2, 355-60 (1959).
RECEIVEDfor review March 13, 1961. Accepted May 5, 1961. Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy,Pittsburgh, Pa., February-March 1961.
